# **Surgical Treatment of Pancreatic Ductal Adenocarcinoma**

Edited by Sohei Satoi Printed Edition of the Special Issue Published in *Cancers*

www.mdpi.com/journal/cancers

## **Surgical Treatment of Pancreatic Ductal Adenocarcinoma**

## **Surgical Treatment of Pancreatic Ductal Adenocarcinoma**

Editor **Sohei Satoi**

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin

*Editor* Sohei Satoi Deparment of Surgery Kansai Medical University Hirakata Japan, and Division of Surgical Oncology University of Colorado Anschutz Medical Campus Aurora USA

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Cancers* (ISSN 2072-6694) (available at: www.mdpi.com/journal/cancers/special issues/pancreatic surgical).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-1866-4 (Hbk) ISBN 978-3-0365-1865-7 (PDF)**

© 2021 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

## **Contents**



Reprinted from: *Cancers* **2021**, *13*, 2031, doi:10.3390/cancers13092031 . . . . . . . . . . . . . . . . **205**

## **About the Editor**

#### **Sohei Satoi**

Dr. Sohei SATOI is a pancreatobiliary surgeon and surgical oncologist with more than 20 years of experience. He is internationally recognized as a leader in the surgical treatment of pancreatic ductal adenocarcinoma including unresectable disease as well as in multimodal treatment of peritoneal dissemination. Dr. Satoi is also keen on developing mitigation strategies for post-pancreatectomy complications. He is active in academic research, with more than 180 peer reviewed papers, and is a member of several scientific journal's editorial boards. His mission is to bring a "cure" to patients with pancreatic ductal adenocarcinoma and to have zero-mortality and less incidence of surgical complication after pancreatectomy.

## **Preface to "Surgical Treatment of Pancreatic Ductal Adenocarcinoma"**

Dear Colleagues,

Surgical resection has been the only chance for a cure in patients with pancreatic ductal adenocarcinoma (PDAC) in recent decades, but the 5-year survival rate is still low (approximately 20%) in PDAC patients who have undergone margin-negative resection only.

A multimodal approach is widely accepted for treating PDAC in the modern era. The implementation of adjuvant chemotherapy or neo-adjuvant treatment has dramatically been enhanced to increase the long-term survival rate of patients with resectable, borderline resectable, and unresectable PDAC in the world. Margin-negative surgical resection still plays a pivotal role in multimodal treatment in patients with PDAC. Therefore, a growing amount of interest has focused on the optimization of the perioperative therapeutic strategy, including regimens of chemo(radio)therapy, the introduction of extended pancreatectomy, and advanced peri-operative management.

This Special Issue highlights the role of surgical resection during multimodal treatments in patients with PDAC from resectable to unresectable diseases to advance our understanding of the surgical treatment of PDAC.

> **Sohei Satoi** *Editor*

### *Editorial* **Surgical Treatment of Pancreatic Ductal Adenocarcinoma**

**Sohei Satoi**

Department of Surgery, Kansai Medical University, Hirakata 573-1010, Japan; satoi@hirakata.kmu.ac.jp

This special issue, "Surgical Treatment of Pancreatic Ductal Adenocarcinoma" contains 13 articles (five original articles, five reviews, and three systematic reviews/meta-analyses) authored by international leaders and surgeons who treat patients with pancreatic ductal adenocarcinoma (PDAC).

Oncological pancreatic surgery requires a deep knowledge of multimodal treatment, accurate preoperative recognition of tumor extension—especially to adjacent major vessels—high-quality technical skills for margin-negative resection, and well-established perioperative management for the reduction of morbidity and mortality. In the modern era, it involves a two-sided advancement toward extended pancreatectomy, such as portal vein or major arterial resection for locally advanced PDAC, as well as minimally invasive surgery for resectable PDAC.

Surgical resection has provided the only chance for a cure in patients with PDAC, but the 5 year survival rate is still low (approximately 20%) in patients with margin-negative resection. The implementation of adjuvant chemotherapy or neoadjuvant therapy has dramatically increased the long term survival of patients with resectable, borderline resectable, and even unresectable PDAC. Margin-negative surgical resection still plays a pivotal role in multimodal treatment in patients with PDAC. Therefore, a growing amount of interest has focused on optimization of the perioperative therapeutic strategy, including multimodal treatment regimens, the introduction of extended pancreatectomy, and advanced perioperative management. Moreover, the introduction of minimally invasive surgery, such as laparoscopic and robotic pancreatic surgery, has been applied worldwide.

This special issue highlights the role of surgical resection in patients with PDAC to advance our understanding of the surgical treatment of PDAC.

With regard to the influence of tumor location on prognosis, among 2483 patients with all types of PDAC, long term survival was significantly better for patients with pancreatic head/uncinate PDAC than with body/tail PDAC, regardless of resectability [1]. Among patients who underwent curative resection, those with head/uncinate cancers had a higher number of T1/T2 tumors, but worse outcomes. Multivariate analysis identified tumor factors, preoperative CA 19-9 level, margin status, and adjuvant therapy, but not tumor location as independent prognostic factors. Margin-negative resection during multimodal treatment is mandatory for long term survival in patients with PDAC.

What can we do to optimize the rate of margin-negative resection? According to the "appropriate dissection range" identified with simulated use of high-quality computed tomography preoperatively, surgeons should carry out "dissection to achieve marginnegative resection", identifying anatomical structures, such as layers, arteries, and veins, as anatomical landmarks to determine the dissection region intraoperatively [2].

Neoadjuvant therapy has been implemented recently to achieve a high proportion of margin-negative resection and negative lymph node metastasis through anatomical and biological shrinkage of borderline resectable tumors. Neoadjuvant therapy followed by surgery, rather than upfront surgery, has been reported to offer clinical benefits to patients with borderline resectable PDAC [3]. Moreover, it is suggested that nutritional management during neoadjuvant therapy may lead to a better prognosis.

Given the multimodal approach with new chemotherapy regimens, such as fluorouracil plus leucovorin, irinotecan, oxaliplatin (FOLFIRINOX), or gemcitabine plus nab-

**Citation:** Satoi, S. Surgical Treatment of Pancreatic Ductal Adenocarcinoma. *Cancers* **2021**, *13*, 4015. https://doi.org/10.3390/cancers 13164015

Received: 22 July 2021 Accepted: 6 August 2021 Published: 10 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

paclitaxel, substantial progress has been made in surgical techniques to address advanced resections [4]. Margin-negative resection in patients with locally advanced PDAC usually requires portal vein or major arterial resection and reconstruction. These mainly vesseloriented technical approaches of pancreatic head resection allow the removal of all putative tumor-infiltrated soft tissue with the utmost aim for an improved R0 resection rate [4].

Aggressive pancreatectomy, such as total pancreatectomy or combined arterial resection, for achieving margin-negative resection has become justified by the principle of total neoadjuvant therapy in recent decades [5]. Further technical standardization and an optimal neoadjuvant strategy are mandatory for the global adoption of aggressive pancreatectomy.

Recently, surgical resection in PDAC has been extended to patients with unresectable PDAC. Additional surgery during multimodal treatment is defined as "conversion surgery" in patients with unresectable PDAC (metastatic and locally advanced disease) which comprises 70–80% of the PDAC population. Although surgical resectability was less than 10% in 398 patients [6] and 469 patients [7] with unresectable PDAC, including metastatic disease, the median survival time after initial treatment was 37 months and 73.7 months in patients who underwent conversion surgery, respectively. The number of candidates for conversion surgery is now increasing with the introduction of modern chemotherapy regimens; however, the actual clinical benefits of resection have not yet been fully investigated. Prospective studies will be needed to explore the clinical benefit of conversion surgery.

A high rate of recurrence, even after margin-negative resection, has been reported in patients with PDAC. Recurrent PDAC, mainly containing liver or peritoneal metastasis, and local recurrence is commonly treated with systemic chemotherapy or best supportive care. The clinical role of surgical resection for patients with isolated local recurrent PDAC after initial pancreatectomy is still under investigation. Although there is a possibility of selection bias, meta-analysis revealed that surgical resection in selected patients with recurrent pancreatic cancer was safe and feasible and might offer a survival advantage [8]. This meta-analysis also suggested that surgery should be considered part of the multimodal management of relapsing pancreatic cancer, and a multidisciplinary approach is essential to choose the most appropriate treatment [8]. Thus, PDAC surgery in the modern era frequently requires extended pancreatectomy; therefore, appropriate patient selection is mandatory. The development of precise biological and anatomical assessments will be urgently needed.

Therefore, novel biomarkers predicting resectability, overall survival, and diseasefree survival should be established promptly. Liquid biopsy involving cancer DNA and circulating tumor cells in the blood may be an additional tool for estimating disease course and outcome in patients with PDAC [9]. Clinical application of liquid biopsy may provide a cancer diagnosis at an earlier stage, enable optimal selection of treatment, and inform prediction of prognosis and recurrence.

While extended pancreatectomy has been developed safely and effectively in patients with locally advanced PDAC, minimally invasive pancreatic surgery has also evolved. Laparoscopic and robotic pancreatectomy are considered safe and feasible for experienced surgeons in well-selected patients with PDAC. With the advancement of minimally invasive techniques and experiences, laparoscopic distal pancreatectomy (LDP) and even laparoscopic pancreaticoduodenectomy (LPD) have been implemented successfully for treating PDAC [10]. However, due to a limited volume of evidence, without doubt, there is a strong need for more high-quality trials to confirm the potential advantages of minimally invasive pancreatic surgery [4].

Optimizing existing pathways for PDAC treatment so that patients realize the benefits of already proven treatments presents a clear opportunity to improve outcomes in the short term. The narrative review [11] focuses on treatments and interventions where there was a clear evidence base to improve outcomes in pancreatic cancer and where there was evidence of variation and undertreatment. The avoidance of preoperative biliary drainage, treatment of pancreatic exocrine insufficiency, prehabilitation and enhanced recovery after surgery, reduction of perioperative complications, optimization of opportunities for elderly patients to receive therapy, optimization of adjuvant chemotherapy, and regular surveillance after surgery are some of the strategies discussed. Each treatment or pathway change represents an opportunity for marginal gain, and the accumulation of marginal gains can result in a considerable benefit to patients. It is essential that surgeons understand that surgery is just one part of a complex pathway and that they are ideally placed to act as change agents to optimize broader pathway improvements.

Other concerns are risk factors for malignancy, defined as high-grade dysplasia and invasive carcinoma in patients with intraductal papillary mucinous neoplasm (IPMN). One meta-analysis revealed risk factors for malignancy as symptoms, size ≥ 3 cm, cystic wall thickening, mural nodule, main pancreatic duct dilatation, abrupt caliber change in the pancreatic duct, lymphadenopathy, elevated carbohydrate antigen 19-9 level, and elevated carcinoembryonic antigen level [12]. Among the above risk factors, the role of main pancreatic duct (MPD) dilatation is important for establishing a simple surgical indication. However, the degree of ductal dilatation that warrants pancreatectomy is still controversial across the existing guidelines. The other meta-analysis concluded that MPD dilatation was an important predictive factor of IPMN malignancy, and 5 mm was a highly sensitive cutoff for the detection of high-risk pre-cancerous or cancerous lesions in resected patients. The need for pancreatectomy should be thoroughly evaluated in patients with ductal dilatations of ≥5 mm for improving surgical patient selection and reducing overall IPMN malignancy mortality [13].

In conclusion, surgical treatment of PDAC has experienced a paradigm shift, from "the only way for cure" in the last century, to "the essential position during multimodal treatment" in the modern era. Pancreatic surgery for PDAC now has two-sided progress. Extended pancreatectomy with vessel resection and reconstruction has been performed safely and effectively in patients with locally advanced PDAC following multimodal treatment. In contrast, the implementation of minimally invasive surgery is also useful in selected patients with PDAC. The establishment of an appropriate surgical indication for predicting an acceptable prognosis is required in the era of multimodal treatment. Biomarkers that inform a surgical indication may be revealed by liquid biopsy in the near future. Sustainable efforts are warranted to establish a role for surgical treatment during multimodal treatment in patients with PDAC who still have high lethality.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


## **The Role of Location of Tumor in the Prognosis of the Pancreatic Cancer**

**Mirang Lee** † **, Wooil Kwon** † **, Hongbeom Kim, Yoonhyeong Byun, Youngmin Han, Jae Seung Kang, Yoo Jin Choi and Jin-Young Jang \***

Department of surgery, Seoul National University Hospital, Seoul 03080, Korea; rang5026@snu.ac.kr (M.L.); willdoc@snu.ac.kr (W.K.); surgeonkhb@snu.ac.kr (H.K.); yoonhyeong@snu.ac.kr (Y.B.); views@snu.ac.kr (Y.H.); 74398@snuh.org (J.S.K.); 74401@snuh.org (Y.J.C.)

**\*** Correspondence: jangjy4@snu.ac.kr; Tel.: +822-2072-2194; Fax: +822-766-3975

† These authors contributed equally to this work as first authors.

Received: 9 June 2020; Accepted: 22 July 2020; Published: 24 July 2020

**Abstract:** Identification of prognostic factors is important to improve treatment outcomes in pancreatic cancer. This study aimed to investigate the effect of the location of pancreatic cancer on survival and to determine whether it was a significant prognostic factor. Altogether, 2483 patients diagnosed with pancreatic cancer were examined. Comparative analysis of clinicopathologic characteristics, survival analysis, and multivariate analysis were performed. Cancers of the pancreatic head or the uncinate process were present in 49.5% of patients. The head/uncinate cancers had more clinical T1/T2 tumors (59.4% vs. 35.5%, *p* < 0.001) and a significantly higher 5-year survival rate (8.9% vs. 7.3%, *p* < 0.001) than the body/tail cancers. The 5-year survival rate in patients with head/uncinate cancers was significantly lower in the resectable (*p* = 0.014) and the locally advanced groups (*p* = 0.007). In patients who underwent resection with curative intent, the 5-year survival rate was lower in the head/uncinate group (*p* = 0.046). The overall outcome of the head/uncinate cancers was better than the body/tail cancers, due to the high proportion of resectable cases. In patients who underwent curative resection, the head/uncinate cancers had a higher number of T1/T2 tumors, but worse outcomes. In the multivariate analysis, tumor location was not an independent prognostic factor for pancreatic cancer.

**Keywords:** pancreatic neoplasm/analysis; pancreatic neoplasm/surgery; tumor location; survival; clinical staging

#### **1. Introduction**

Pancreatic cancer is one of the leading causes of cancer-related mortality in developed countries and one of the most lethal malignant neoplasms worldwide [1]. Its prognosis might be poor, and accurate prediction of the prognosis is important for patients as well as clinicians in the management of pancreatic cancer.

Surgical approach to pancreatic cancer and its prognosis greatly differ according to the tumor location [2–4]. Some authors have argued that pancreatic body and tail cancers have a worse prognosis due to delayed diagnosis. Others have reported that according to the tumor stage at diagnosis, pancreatic body and tail cancers showed superior survival than pancreatic head cancers, in localized and resectable tumors. Despite these differences, tumor location was never taken into consideration in any edition of the American Joint Committee on Cancer (AJCC) staging system, since the first edition in 1978. Thus, the effect of location on pancreatic cancer needs to be highlighted.

Several issues related to tumor location need to be scrutinized in depth. One of them is to clarify whether tumor location affects the prognosis of pancreatic cancer and if it does, the manner in which it affects the prognosis. Furthermore, it should be examined whether tumor location affects the prognosis

to such an extent that it should be reflected in the staging system of pancreatic cancer. With these questions in mind, the present study aimed to compare the survival outcomes and clinicopathological features of pancreatic cancer, according to its location.

#### **2. Results**

#### *2.1. Patient Demographics and Survival Outcomes*

Altogether, 2483 patients were identified. Among these, 1228 patients (49.5%) had tumors in the pancreatic head or the uncinate process (PHU group) and 1255 patients (50.5%) had tumors in the pancreatic body or the tail (PBT group). Demographics and clinicopathological features are summarized in Table 1. The mean age was comparable between the PHU and the PBT groups (64.3 years and 64.0 years, respectively *p* = 0.468). The sex ratio was also similar between the groups, showing male predominance (1:0.68 and 1:0.68, respectively; *p* = 0.097).


**Table 1.** Demographics and clinicopathological features of overall patients.

PHU—tumors in the pancreas head or uncinated process; PBT—tumors in the pancreas body and tail. Continuous variables were expressed as median (range). Statistical significance when *p* value < 0.05.

Mean tumor size was significantly different between the PHU group and the PBT group (3.4 cm and 4.3 cm, respectively; *p* < 0.001). The proportion of clinical T stages was significantly different (*p* < 0.001). The PHU group had a higher proportion of cT2 (50.2% in PHU vs. 29.2% in PBT) tumors. The PBT group had a higher proportion of cT3 and cT4 tumors than the PHU group (cT3: 27.3% vs. 12.4% and cT4: 37.1% vs. 28.3%, respectively).

The proportion of tumors was significantly different in terms of classification according to resectability between the PHU and the PBT groups (*p* < 0.001). The PHU group had a higher proportion of resectable and borderline resectable pancreatic cancers (resectable—36.6% vs. 18.2% and borderline resectable: 6.9% vs. 3.1%, respectively) and a lower proportion of metastatic pancreatic cancer (31.0% vs. 54.7%, respectively) than the PBT group. The proportion of locally advanced pancreatic cancers was similar between the groups (25.5% and 24.1% in the PHU and the PBT groups, respectively).

The median survival in all patients was 11 months and the 5-year survival rate was 8.1%. The PHU group demonstrated significantly better survival than the PBT group (median survival—12 vs. 10 months, and 5-year survival—8.9% vs. 7.3%, respectively; *p* < 0.001) (Figure 1).

respectively).

The proportion of tumors was significantly different in terms of classification according to resectability between the PHU and the PBT groups (*p* < 0.001). The PHU group had a higher proportion of resectable and borderline resectable pancreatic cancers (resectable—36.6% vs. 18.2% and borderline resectable: 6.9% vs. 3.1%, respectively) and a lower proportion of metastatic pancreatic cancer (31.0% vs. 54.7%, respectively) than the PBT group. The proportion of locally advanced pancreatic cancers was similar between the groups (25.5% and 24.1% in the PHU and the PBT groups,

The median survival in all patients was 11 months and the 5-year survival rate was 8.1%. The

10 months, and 5-year survival—8.9% vs. 7.3%, respectively; *p* < 0.001) (Figure 1).

**Figure 1.** The survival curves of pancreatic cancer in the head/uncinate region and body/tail regions in all patients are illustrated. **Figure 1.** The survival curves of pancreatic cancer in the head/uncinate region and body/tail regions in all patients are illustrated.

#### *2.2. Demographics of the Patients Who Underwent Resection 2.2. Demographics of the Patients Who Underwent Resection*

Among 705 patients who were advised to undergo curative resection, 28 patients who underwent neoadjuvant treatment and 31 patients who ended up having non-curative surgery were excluded. Thus, 646 patients underwent curative resection. Altogether, 432 (66.9%) patients in the PHU group and 214 (33.1%) patients in the PBT group underwent curative resection. The PHU group had a significantly smaller tumor size, more angiolymphatic invasion and perineural invasion, a lower proportion of T3 and T4 tumors, a higher proportion of N2 and a lower proportion of N0 tumors, greater recurrence, and lower incidence of systemic recurrence, when compared to the PBT group. There were no differences in carcinoembryonic antigen and carbohydrate antigen (CA) 19-9 levels, lymph node (LN) metastasis rate, and the proportion of patients who received adjuvant therapy. Demographics and clinicopathological features are summarized in Table 2. Among 705 patients who were advised to undergo curative resection, 28 patients who underwent neoadjuvant treatment and 31 patients who ended up having non-curative surgery were excluded. Thus, 646 patients underwent curative resection. Altogether, 432 (66.9%) patients in the PHU group and 214 (33.1%) patients in the PBT group underwent curative resection. The PHU group had a significantly smaller tumor size, more angiolymphatic invasion and perineural invasion, a lower proportion of T3 and T4 tumors, a higher proportion of N2 and a lower proportion of N0 tumors, greater recurrence, and lower incidence of systemic recurrence, when compared to the PBT group. There were no differences in carcinoembryonic antigen and carbohydrate antigen (CA) 19-9 levels, lymph node (LN) metastasis rate, and the proportion of patients who received adjuvant therapy. Demographics and clinicopathological features are summarized in Table 2.


**Table 2.** Demographics and clinicopathological features of resected patients.


**Table 2.** *Cont*.

PHU—tumors in the pancreas head or uncinated process; PBT—tumors in the pancreas body and tail; PPPD—pylorus-preserving pancreatoduodenectomy. Continuous variables were expressed as median (range). Statistically significant when *p* value < 0.05.

#### *2.3. Survival Analysis of the Patients Who Underwent Resection*

The median survival duration was 25 months and the 5-year survival was 23.6%. For the PHU group, the median survival duration was 23 months and the 5-year survival was 20.8%. For the PBT group, the median survival duration was 30 months and the 5-year survival was 29.7%. Thus, the survival outcome in the PBT group was significantly superior to that in the PHU group (*p* = 0.046) (Figure 2).

(Figure 2).

The median survival duration was 25 months and the 5-year survival was 23.6%. For the PHU group, the median survival duration was 23 months and the 5-year survival was 20.8%. For the PBT group, the median survival duration was 30 months and the 5-year survival was 29.7%. Thus, the

**Figure 2.** The survival curves of pancreatic cancer in the head/uncinate region and body/tail regions in resected patients are illustrated. **Figure 2.** The survival curves of pancreatic cancer in the head/uncinate region and body/tail regions in resected patients are illustrated.

Survival outcomes were compared according to the T category. For T1, T2, and T4 tumors, the PHU group had worse outcomes compared to the PBT group. The difference was not significant for the T1 (median survival 34 vs. 41 months, respectively; *p* = 0.288) and T4 tumors (median survival 6 vs. 8 months, respectively; *p* = 0.067). A significant difference was found in T2 tumors with a median survival of 22 months for the PHU group (5-year survival 19.4%) and a median survival of 34 months for the PHU group (5-year survival 34.8%) (*p* = 0.005, Figure 3). Survival outcomes were compared according to the T category. For T1, T2, and T4 tumors, the PHU group had worse outcomes compared to the PBT group. The difference was not significant for the T1 (median survival 34 vs. 41 months, respectively; *p* = 0.288) and T4 tumors (median survival 6 vs. 8 months, respectively; *p* = 0.067). A significant difference was found in T2 tumors with a median survival of 22 months for the PHU group (5-year survival 19.4%) and a median survival of 34 months *Cancers*  for the PHU group (5-year survival 34.8%) ( **2020** *p* = 0.005, Figure 3). , *12*, x 6 of 14

**Figure 3.** The survival curves of pancreas head/uncinate cancer and pancreas body cancer according to the T categories. **Figure 3.** The survival curves of pancreas head/uncinate cancer and pancreas body cancer according to the T categories.

positive disease, the PHU group had worse outcomes than the PBT group, but the difference lacked

According to the prognostic groups of the AJCC cancer staging system (edition 8), there were no differences in survival outcomes between the PHU and the PBT groups, in all stages from stage Ia to

Tumor location, histological grade, margin status, angiolymphatic invasion, venous invasion, perineural invasion, T category, N category, adjuvant chemotherapy, adjuvant radiotherapy, and preoperative CA 19-9 were significantly associated with survival. In the multivariate analysis, tumor location did not reach statistical significance (vs. PBT: hazard ratio [HR] 1.174, confidence interval [CI] 0.932–1.478, *p* = 0.173). Histological grade, margin status, angiolymphatic invasion, venous invasion, T4 stage, lymph node metastasis, adjuvant chemotherapy, adjuvant radiotherapy, and

**Table 3.** Univariate and multivariate analysis comparing the 5-year survival rates in resected patients.

**variables Univariate Multivariate** 

≥65 345 23.6 0.070 1.095 0.881–1.362 0.414

Body, tail 214 29.7 0.046 1.174 0.932–1.478 0.173

*n* **5 YSR, %** *p* **value HR 95%CI** *p* **value** 

statistical significance (19 vs. 25 months, respectively; *p* = 0.112).

preoperative CA 19-9 were independent prognostic factors (Table 3).

Female 266 27.3 0.269

Male 380 21.2

<65 301 24.0

Head 432 20.8

stage III (Figure S1).

Sex

Age(years)

Site of tumor

Complication

*2.4. Prognostic Factors of Pancreatic Cancer* 

In node-negative disease, the PHU group had worse median survival than the PBT group, but the difference was not significant (33 vs. 39 months, respectively; *p* = 0.454). Similarly, in the node-positive disease, the PHU group had worse outcomes than the PBT group, but the difference lacked statistical significance (19 vs. 25 months, respectively; *p* = 0.112).

According to the prognostic groups of the AJCC cancer staging system (edition 8), there were no differences in survival outcomes between the PHU and the PBT groups, in all stages from stage Ia to stage III (Figure S1).

#### *2.4. Prognostic Factors of Pancreatic Cancer*

Tumor location, histological grade, margin status, angiolymphatic invasion, venous invasion, perineural invasion, T category, N category, adjuvant chemotherapy, adjuvant radiotherapy, and preoperative CA 19-9 were significantly associated with survival. In the multivariate analysis, tumor location did not reach statistical significance (vs. PBT: hazard ratio [HR] 1.174, confidence interval [CI] 0.932–1.478, *p* = 0.173). Histological grade, margin status, angiolymphatic invasion, venous invasion, T4 stage, lymph node metastasis, adjuvant chemotherapy, adjuvant radiotherapy, and preoperative CA 19-9 were independent prognostic factors (Table 3).


**Table 3.** Univariate and multivariate analysis comparing the 5-year survival rates in resected patients.


**Table 3.** *Cont*.

YSR—year survival rate; HR—hazard ratio; CI—confidence interval. The variables with *p*-value less than 0.1 in univariate analysis were included in the multivariate analysis.

When analyzed separately for the PHU and the PBT groups, factors associated with survival in the univariate analysis were similar between the groups and similar to the factors associated with the overall patient population. For the PHU group, all associated categories were similar to those associated with the overall patient population. For the PBT group, T2 stage and preoperative CA 19-9 were not associated with survival, while age was associated with survival, when compared to the overall patient population.

Multivariate analysis showed that poorly differentiated histological grade, angiolymphatic invasion, perineural invasion, T4 stage, N2 stage, adjuvant chemotherapy, adjuvant radiotherapy, and preoperative CA 19-9 were significantly associated with survival in the PHU group. In the PBT group, histological grade, margin status, venous invasion, and adjuvant chemotherapy were significantly associated with survival (Table 4).


**Table 4.** Comparison of independent risk factors of pancreatic cancer in PHU and PBT.

*Cancers* **2020**, *12*, 2036




13

#### **3. Discussion**

The AJCC staging system was revised for the eighth time since its first edition in 2018. Its validity was demonstrated in several studies [5–7]. The AJCC staging system always considered pancreatic cancers in terms of the whole pancreas, without dividing the pancreas according to the location, since pancreatic cancers in the head/uncinate process and those in the body/tail share the same prognosis and have comparable tumor biology. However, pancreatic cancer is usually treated according to the location. Many studies investigated pancreatic cancers separately according to the location [2,4,8–12]. Furthermore, many studies investigated whether the subjects underwent distal pancreatectomy or pancreatoduodenectomy, which is a reflection of the location of the tumor [13–16]. Pancreatic cancer is often not looked at somewhat differently. In this light, it must be clarified whether pancreatic head cancers and pancreatic body or tail cancers have comparable outcomes and oncological behaviors.

Traditionally, pancreatic cancers in the body/tail are believed to have a worse prognosis compared to pancreatic head cancers. This finding was supported by several studies [2,8–10,17] and it was also reproduced in the present study. The 5-year survival percentages and the median survival durations were significantly better for the PHU group than for the PBT group, in all pancreatic cancers, regardless of their resectability. The poor outcome of pancreatic cancers in the body/tail is usually explained by their late detection.

While a pancreatic head cancer might cause obstructive jaundice as the tumor progresses, patients with pancreatic body/tail cancers do not show symptoms until the tumor size increases sufficiently to cause abdominal pain and colon obstruction. In the present cohort, the tumor size measured on the cross-sectional images was significantly greater in the PBT group. Larger tumors reduce the possibility of resectability, which is also reflected in the results of the present study. In the present study, 36.6% of the pancreatic head cancers were deemed resectable, while only 18.2% of the pancreatic body/tail cancers were deemed resectable.

Late detection of the pancreatic body and tail tumors allows them to grow, reducing their resectability. It also increases the possibility of systemic involvement. Other studies that investigated pancreatic cancers according to their locations showed that pancreatic body and tail cancers often present with distant metastases at the time of diagnosis [2,11]. The present study also confirmed a higher proportion of systemic spread at presentation (54.7% in the body/tail cancers and 31.0% in the head/uncinate region cancers).

A completely opposite set of findings was observed when only the resected cases were considered. In the resected cases, pancreatic cancers in the head/uncinate region demonstrated significantly worse survival than those in the body/tail region. Many studies found similar results in resectable pancreatic cancers in the head/uncinate regions [11,12,18], while some studies failed to show worse results for the head region when compared to the body/tail region [9,10,13–16,19]. However, none of these studies showed significantly worse outcomes in pancreatic body/tail cancers [12].

Studies that demonstrated comparable outcomes between resectable pancreatic cancers in the head and those in the body/tail should be noted for their study population. Studies conducted by Sohn et al. [13], Wade et al. [14], and Brennan et al. [19] published in 2000, 1995, and 1996, respectively, are considered the historic ones. Their study populations were collected from as early as 1984 and up to 1999. During this period, safety and oncological feasibility of pancreatic cancer surgery was more of an issue. Furthermore, adjuvant treatment, which is currently an important part of pancreatic cancer treatment, was not established. The studies from the late 2000s and the 2010s had similar problems regarding patient populations as those associated with the patient populations from the 1980s and the 1990s, even though they included more recent cohorts [9,15,16].

Only one study that included 351 patients showed superior outcomes in the resected pancreatic head cancers, when compared with the resected body/tail cancers [4]. The median survivals of patients with pancreatic head cancers and of those with pancreatic body/tail cancers were 16 and 11 months, respectively. This rather poor survival outcome in patients with resected tumors limited the value of this study. All the other studies reported comparable or superior outcomes in resected pancreatic

body/tail cancers than in resected pancreatic head cancers. Therefore, based on the recent literature and the results of the present study, it could be safely concluded that resected pancreatic cancers in the body/tail region have better outcomes than those in the head region. As such is the case, resection of pancreatic body/tail cancers should not be discouraged because of the poor overall prognosis, but rather should be attempted, whenever deemed resectable.

When analyzed according to the T stages, significant difference was observed in survival between the groups for T2 tumors. The PHU group showed worse outcomes than the PBT group for T1 tumors, but the difference was not significant. This finding might perhaps be attributed to small-sized subgroups. Thus, earlier T categories (T1 and T2) demonstrated significantly worse median survival and 5-year survival (24 months and 21.8%) in the PHU group than in the PBT group (34 months and 37.2%) (*p* = 0.003). Meng et al. [12] also found that resected pancreatic head cancers had worse outcomes in the earlier T stages, but significant difference was observed only for the T1 stage.

When stratified according to the N stage and the prognostic groups, the survival rates were not significantly different. There was a tendency toward worse survival for pancreatic head/uncinate cancers in the N0 and LN metastasis groups. For the prognostic groups, pancreatic head/uncinate cancers in stages IB, IIB, and III tended to have worse survival. Re-evaluation using a larger cohort or meta-analysis might clarify the effect of cancer location in each stratified analysis.

There were varying results regarding whether the location of pancreatic cancer was an independent prognostic factor. The present study found that cancer location was not an independent risk factor (head vs. body/tail: HR 1.174, CI 0.932–1.478, *p* = 0.173). Similarly, Ruess et al. [16] and van Erning et al. [10] did not identify location as an independent risk factor. Several other studies suggested that location was a significant risk factor [2,4,9,11,12,18]. Therefore, the status of cancer location as an independent prognostic factor is still controversial and needs further high-level evidence.

The difference in survival outcomes between the locations might be due to plain anatomical differences causing symptoms at different time intervals. There might be additional differences in tumor biology and behavior. To investigate the differences in tumor biology and behavior, clinicopathological features of pancreatic cancers in the head/uncinate process and cancers in the body/tail region were analyzed and compared. Some differences were present, but common risk factors were also observed. Hence, the results are unclear and a definite conclusion cannot be obtained.

Additionally, differences on genetic and molecular levels should also be considered. The present study did not examine this aspect, but previous studies examined genetic profiles. Birnbaum et al. [20] found differences in 334-gene expression signature between tumors in the head and those in the body/tail. Dreyer et al. [8] reported that tumors might have different molecular pathology, according to their location and the body/tail tumors are enriched with gene programs involved in tumor invasion, epithelial-to-mesenchymal transition, and poor antitumor immune response. This is an important area of research, as the differences on genetic and molecular levels might open a new era of more tailored treatment approaches, according to the location.

The present study had some limitations. The study was retrospective in nature. In addition, the overall patient dataset was acquired through a clinical data warehouse. Hence, more specific variables could not be retrieved in detail. As the present study was performed at a tertiary hospital, many patients visited after being diagnosed at other primary or secondary hospitals, which might have resulted in bias regarding the date of diagnosis. The study population was insufficient for subgroup analyses after stratification.

#### **4. Materials and Methods**

#### *4.1. Study Design*

The study was approved by the ethical committee of the Institutional Review Board of Seoul National University Hospital (IRB No. H-1902-012-1006). Seoul National University Hospital's Clinical Data Warehouse was searched for patients who were diagnosed with pancreatic ductal adenocarcinoma between 2005 and 2016. A retrospective cohort study was performed.

This research was supported by the Collaborative Genome Program for Fostering New Post-Genome Industry of the National Research Foundation funded by the Ministry of Science and ICT (NRF-2017M3C9A5031591), and by a grant from the Korean Health Technology R and D Project, Ministry of Health and Welfare, Republic of Korea (HI14C2640).

#### *4.2. Patient Selection*

After the identification of patients with pancreatic cancer from the database, those with multiple tumors in both the head and the body/tail were excluded. Patients who had tumors across the junction of the head and the body were also excluded, as grouping according to the location was ambiguous in these tumors. Data regarding age, sex, tumor location, tumor size on radiological images, clinical feature (T) classification, and classification based on resectability were collected.

Further subgroup analysis was performed for patients who underwent resection with curative intent. Among all patients, 646 patients who underwent resection of pancreatic cancer with curative intent were examined. Patients who underwent only palliative operation including bypass or open biopsy were excluded. Since neoadjuvant treatment can alter the final pathological staging, patients who received neoadjuvant therapy were also excluded. Detailed information about the demographic and clinicopathological factors of these patients was obtained through a thorough review of their electronic medical records.

#### *4.3. Determination of Tumor Location and Clinical T Staging*

Computed tomography (CT) or magnetic resonance imaging (MRI) records of all patients were reviewed. An imaginary tangential line over the left border of the superior mesenteric vein or the portal vein was drawn on the CT image. The head/uncinate pancreatic cancer group (PHU) was defined as patients with tumors on the right side of this line. The body/tail pancreatic cancer group (PBT) was defined as patients with tumors on the left side of this line.

Clinical T staging was performed according to the AJCC staging system (edition 8) for pancreatic cancer. Tumor size was measured using CT and MRI.

#### *4.4. Definition of Survival and Data Collection*

Overall survival was used for the analysis. It was defined as the interval between the date of diagnosis and the date of death or the last follow-up. Survival status was acquired from the Ministry of Interior and Safety of Korea. Patients who were alive on 20 February 2019 were censored.

#### *4.5. Statistical Analysis*

Fisher's exact test and chi-squared test were used to compare categorical variables and unpaired two-sided Student's t-test was used to compare continuous variables between patients with tumors located in the head/uncinate process and patients with tumors in the body/tail. The Kaplan-Meier method with log-rank test was used for survival analysis. Cox regression test was used for the univariate and the multivariate analyses. A *p*-value < 0.050 was considered to be statistically significant. IBM SPSS statistics for Windows version 24 (IBM Corp., Armonk, NY, USA) was used for statistical analyses.

#### **5. Conclusions**

The prognosis of pancreatic cancers differed according to the location of the tumors. Pancreatic head cancers showed a better overall prognosis than pancreatic body/tail cancers, which might be related to a higher proportion of systemic involvement in the latter. On the contrary, resected pancreatic head cancers showed a worse prognosis than resected pancreatic body/tail cancers, especially in the earlier T stages. Tumor location was not an independent risk factor for pancreatic cancer.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/12/8/2036/s1. Figure S1: The survival curves of pancreas head/uncinate cancer and pancreas body cancer in terms of the prognostic group, according to the eighth edition of AJCC cancer staging system.

**Author Contributions:** Conceptualization, W.K. and J.-Y.J.; Data curation, M.L., W.K., H.K. and Y.H.; Formal analysis, M.L.; Funding acquisition, Y.H.; Investigation, M.L., J.S.K. and Y.J.C.; Methodology, Y.B.; Project administration, W.K., H.K. and J.-Y.J.; Resources, J.S.K. and Y.J.C.; Supervision, W.K.; Validation, W.K., H.K., Y.B., Y.H. and J.-Y.J.; Writing—original draft, M.L.; Writing—review & editing, M.L., W.K., Y.B. and J.-Y.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by the Collaborative Genome Program for Fostering New Post-Genome Industry of the National Research Foundation, funded by the Ministry of Science and ICT (NRF-2017M3C9A5031591) and the Korean Health Technology R&D Project, Ministry of Health and Welfare (HI14C2640), Korea.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Reconsideration of the Appropriate Dissection Range Based on Japanese Anatomical Classification for Resectable Pancreatic Head Cancer in the Era of Multimodal Treatment**

**Yuichi Nagakawa 1,\* , Naoya Nakagawa <sup>1</sup> , Chie Takishita <sup>1</sup> , Ichiro Uyama <sup>2</sup> , Shingo Kozono <sup>1</sup> , Hiroaki Osakabe <sup>1</sup> , Kenta Suzuki <sup>1</sup> , Nobuhiko Nakagawa <sup>1</sup> , Yuichi Hosokawa <sup>1</sup> , Tomoki Shirota <sup>1</sup> , Masayuki Honda <sup>1</sup> , Tesshi Yamada 1,3 , Kenji Katsumata <sup>1</sup> and Akihiko Tsuchida <sup>1</sup>**


**Simple Summary:** Although the survival benefit of "regional lymph node dissection" for pancreatic head cancer remains unclear, the R0 resection rate is reportedly associated with prognosis. We reviewed the literature that could be helpful in determining the appropriate resection range. The recent development of high-quality computed tomography has made it possible to evaluate the extent of cancer infiltration. Even if the "dissection to achieve R0 resection" range is simulated based on the computed tomography evaluation, it is difficult to identify the range intraoperatively. It is necessary to be aware of the anatomical landmarks to determine the appropriate dissection range intraoperatively.

**Abstract:** Patients with resectable pancreatic cancer are considered to already have micro-distant metastasis, because most of the recurrence patterns postoperatively are distant metastases. Multimodal treatment dramatically improves prognosis; thus, micro-distant metastasis is considered to be controlled by chemotherapy. The survival benefit of "regional lymph node dissection" for pancreatic head cancer remains unclear. We reviewed the literature that could be helpful in determining the appropriate resection range. Regional lymph nodes with no suspected metastases on preoperative imaging may become areas treated with preoperative and postoperative adjuvant chemotherapy. Many studies have reported that the R0 resection rate is associated with prognosis. Thus, "dissection to achieve R0 resection" is required. The recent development of high-quality computed tomography has made it possible to evaluate the extent of cancer infiltration. Therefore, it is possible to simulate the dissection range to achieve R0 resection preoperatively. However, it is often difficult to distinguish between areas of inflammatory changes and cancer infiltration during resection. Even if the "dissection to achieve R0 resection" range is simulated based on the computed tomography evaluation, it is difficult to identify the range intraoperatively. It is necessary to be aware of anatomical landmarks to determine the appropriate dissection range during surgery.

**Keywords:** pancreatic cancer; pancreaticoduodenectomy; mesopancreas; superior mesenteric artery; nerve and fibrous tissues; adjuvant chemotherapy; lymph node dissection; R0 resection

**Citation:** Nagakawa, Y.; Nakagawa, N.; Takishita, C.; Uyama, I.; Kozono, S.; Osakabe, H.; Suzuki, K.; Nakagawa, N.; Hosokawa, Y.; Shirota, T.; et al. Reconsideration of the Appropriate Dissection Range Based on Japanese Anatomical Classification for Resectable Pancreatic Head Cancer in the Era of Multimodal Treatment. *Cancers* **2021**, *13*, 3605. https://doi.org/10.3390/ cancers13143605

Academic Editor: Mathias Worni

Received: 14 June 2021 Accepted: 15 July 2021 Published: 19 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Pancreatic ductal adenocarcinoma (PDAC) is recognized as having one of the poorest prognoses of all tumors. Resection is the only treatment that can result in long-term survival. Several randomized controlled trials have shown that extended lymph node dissection does not provide survival benefits in patients with pancreatic head cancer, despite a prolonged operative time and increased blood loss [1–5]. Regional lymph node dissection for pancreatic head cancer has been performed in many facilities, but its survival benefit remains unclear. On the other hand, pancreatic cancer treatment has dramatically changed recently owing to the development of effective chemotherapy. Adjuvant chemotherapy is essential for improving the prognosis of pancreatic cancer [6,7]. A randomized prospective study showed that the introduction of preoperative chemotherapy led to a prolonged prognosis in patients with pancreatic cancer [8]. Most of the recurrence patterns of resectable pancreatic cancer are distant metastases, and resectable pancreatic cancer is considered to be a systemic disease with micrometastasis. Thus, multimodal treatment is required to improve the prognosis of resectable pancreatic cancer, and pancreatic resection should be performed with consideration of preoperative and postoperative treatment.

However, many studies reported that the R0 resection rate is associated with prognosis. The recent development of high-quality high-resolution multi-detector computed tomography (MDCT) has made it possible to evaluate the extent of cancer progression, which makes it possible to simulate the appropriate dissection range to achieve R0 surgery before surgery. Even if the dissection range is simulated preoperatively, an accurate understanding of the anatomical structure is required to identify the dissection range during surgery. In the era of multidisciplinary treatment for resectable pancreatic cancer, we reviewed the literature that could be helpful in determining the appropriate resection range.

#### **2. Is "Regional LYMPH Node Dissection" Required?**

Patients with pancreatic cancer often have lymph node metastasis, and many studies have reported that lymph node metastasis is a prognostic factor [9–11]. Prior to the development of effective adjuvant chemotherapy, extended lymph node dissection, including para-aortic lymph nodes, was performed to prevent local recurrence [12,13]. However, several randomized controlled trials have shown that extended lymphadenectomy does not provide survival benefits in patients with pancreatic head cancer, despite a prolonged operative time and increased blood loss [1–5]. However, it has been reported that the number of retrieved lymph nodes is associated with R0 resection rates and survival [14].

The regional lymph nodes are numbered according to the Japanese Pancreatic Cancer classification [15]. Regional lymph nodes for pancreaticoduodenectomy are classified into Group 1 (8a, 8p, 13a, 13b, 17a and 17b) and Group 2 (5.6, 12a, 12b, 12p, 14p and 14d). The lymph node dissection in Group 1 is defined as D1 dissection, and the lymph node dissection of Group 1 and Group 2 is defined as D2 dissection (Figure 1A). However, it was reported that there was no significant difference in prognosis between D1 and D2 dissections in a randomized controlled trial [5] (Table 1), and it is still debated whether prophylactic dissection of regional lymph nodes improves prognosis [16]. Using surgical results of 495 patients with PDAC, Imamura et al. calculated the efficacy index for each lymph node station by multiplying the frequency of lymph node metastasis to the station and survival to clarify the optimal extent of lymph node dissection. Their results indicated that the efficacy of lymph node dissection differs between uncinate process cancer and pancreatic neck cancer, and the extent of dissection should be determined according to the location of the tumor. They also showed that the site of regional lymph node and lymph node recurrence pattern are different, indicating that it may be necessary to reconsider the need for regional lymph node dissection [15].

**Figure 1.** (**A**) The regional lymph nodes are numbered according to the Japanese Pancreatic Cancer classification. Green, D1 region, purple, D2 region. (**B**) Extra-pancreatic nerve plexus around the SMA nerve plexus in the Japanese pancreatic cancer classification. (**C**) Intensive NFTs spreading around the SMA are classified into four areas. Nagakawa et al. [16] classified "intensive NFTs" around the pancreatic head into areas A–D. They also found the three SMA regions (SMAI-III) that can be easily exposed. These regions become anatomical landmarks as "dissection-guiding points" to uniformly dissect each area A–D. PLphI, pancreatic head nerve plexus I; PLphII, pancreatic head nerve plexus II; CA, celiac artery; CHA, common hepatic artery; SMA, superior mesenteric artery; IPDA, inferior pancreaticoduodenal artery; J1A, first jejunal artery; J2A, second jejunal artery; PDJV, proximal dorsal jejunal vein. **Figure 1.** (**A**) The regional lymph nodes are numbered according to the Japanese Pancreatic Cancer classification. Green, D1 region, purple, D2 region. (**B**) Extra-pancreatic nerve plexus around the SMA nerve plexus in the Japanese pancreatic cancer classification. (**C**) Intensive NFTs spreading around the SMA are classified into four areas. Nagakawa et al. [16] classified "intensive NFTs" around the pancreatic head into areas A–D. They also found the three SMA regions (SMAI-III) that can be easily exposed. These regions become anatomical landmarks as "dissection-guiding points" to uniformly dissect each area A–D. PLphI, pancreatic head nerve plexus I; PLphII, pancreatic head nerve plexus II; CA, celiac artery; CHA, common hepatic artery; SMA, superior mesenteric artery; IPDA, inferior pancreaticoduodenal artery; J1A, first jejunal artery; J2A, second jejunal artery; PDJV, proximal dorsal jejunal vein.


**Table 1.** Dissection area in randomized controlled trials of extended lymph node dissection and standard dissection for pancreatic head cancer.

\*: Lymph node numbers are listed according to the Japanese Pancreatic Cancer classification. MST: median survival time.

In the area around the superior mesenteric artery (SMA), "regional lymph node dissection" also dissects the adipose and connective tissues around the regional lymph nodes, which is almost the same dissection range as "dissection to achieve R0 resection." On the other hand, the 14p, 14d, and 8p lymph nodes, which are located around the CHA and the SMA, cannot be identified during resection because the lymph nodes are covered with many nerves and fibers. These regional lymph nodes may be confused with other numbers of lymph nodes. Thus, it is difficult to identify the precise location of each regional lymph node during surgery. Novel criteria may be needed to determine the appropriate lymph node dissection area [17].

Here, it should be noted that "regional lymph node dissection" and "dissection to achieve R0 resection" have different purposes. "Dissection to achieve R0 resection" is performed to avoid residual cancer infiltration, whereas "regional lymph node dissection" is performed to prevent recurrence of the lymph nodes. Thus, "regional lymph node dissection" and "dissection to achieve R0 resection" should be separately when considering the appropriate dissection range for resectable PDAC. Multimodal treatment, including neoadjuvant therapy and postoperative adjuvant chemotherapy, dramatically improve prognosis. Regional lymph nodes with no suspected metastases on preoperative imaging may become areas treated with preoperative and postoperative adjuvant chemotherapy. Further discussion is needed to clarify the necessity of "regional lymph node dissection."

#### **3. Is "Dissection to Achieve R0 RESECTION" Required?**

Many studies have described the need for R0 resection to achieve long-term survival, and the results of most studies have shown that R0 resection improves the survival rate of patients with resectable PDAC who have undergone pancreaticoduodenectomy (PD) [18,19]. Ghaneh et al. [20] analyzed data from the European Study Group for Pancreatic Cancer-3 randomized controlled trial and found that R1 (direct) resections were associated with significantly reduced overall and recurrence-free survival following pancre-

atic cancer resection. Resection margin involvement was also associated with an increased risk of local recurrence. Based on these results, the National Comprehensive Cancer Network (NCCN) guidelines have described that the goals of surgical extirpation of pancreatic carcinoma focus on the achievement of an R0 resection, as a margin-positive specimen is associated with poor long-term survival. In contrast, Schmocke et al. [21] retrospectively examined 468 patients with resectable pancreatic cancer or borderline resectable pancreatic cancer who received preoperative treatment. They reported that margin status was not a significant predictor of overall survival or relapse-free survival in multivariate analysis, but the clinical stage, duration of *N*-acetyl cysteine treatment, nodal status, histopathologic treatment response score, and receipt of adjuvant chemotherapy were factors associated with overall survival. In contrast, in pancreaticoduodenectomy with a complicated cutting surface, the R0 resection rate may differ depending on the evaluation and slicing methods [18]. Additionally, the inking of the cut surface according to a defined color code leads to an accurate R0/R1 evaluation [22]. Two definitions have been reported in assessing R1 [23]. American and Japanese classifications define R1 as direct microscopic involvement at the resection margin (0 mm rule) [15,24], and the Royal University of Pathologists classification defines R1 as the presence of cancer cells within 1 mm of the resection margin (1 mm rule) [25]. It is still unclear which classification reflects the prognosis [26]. To clarify the need to achieve R0 resection to prolong the prognosis in the era of multimodal treatments, the pathological evaluation should be standardized. Currently, there is little evidence that "R0 resection is not needed" to improve prognosis. Therefore, even in the era of multimodal treatment, resectable pancreatic cancer may require surgery to achieve R0 resection.

#### **4. The Issue Regarding Tumor Infiltration of Nerve and Fibrous Tissues**

Dense connective tissues exist around the pancreatic head, which is composed of intensive nerve and fibrous tissues (NFTs). It has been reported that dissection of the NFTs around the pancreatic head is important for achieving R0 resection because PDAC often infiltrate these NFTs [27,28]. However, the appropriate dissection range of NFTs has not been fully discussed, and a common classification showing the anatomical structure of NFTs around the pancreatic head is needed to determine the dissection range. Several classifications have been shown for the anatomy of NFTs. The Japanese classification for pancreatic cancer shows the anatomy of NFTs around the pancreatic head. In this classification, the major NFTs connecting to the pancreatic head are classified into two pathways. One is the pathway from the right celiac ganglion to the posterior side of the pancreas head (pancreatic head plexus I; PLph I), and the other is the pathway from the SMA nerve plexus to the left side of the uncinate process (pancreatic head nerve plexus II; PLph II) [15]. Nagakawa et al. [29] classified the intensive NTFs spreading around the SMA into four areas based on the autopsy findings. Area A: NFTs spreading from the right celiac ganglion and the superior side of the pancreatic head and the posterior side of the hepatoduodenal ligament. Area B: NFTs spreading from the SMA nerve plexus and the uncinate process. Area C: NFTs spreading from the SMA nerve plexus to the anterior side of the jejunal mesentery. Area D: NFTs spreading from the inferior side of the uncinate process to the posterior side of the jejunal mesentery. They also found three SMA nerve plexus regions without branching nerves (SMA I-III) and described that these regions become good anatomical landmarks to identify the SMA nerve plexus before stating these NFTs areas. These anatomical classifications may become good criteria for determining the appropriate dissection range of NFTs.

#### **5. Determination of the Appropriate Dissection Range**

Intraoperative pathological diagnosis using frozen section is generally performed to determine the pancreatic cutting line to avoid positive pancreatic neck margins. Additionally, resectability status is also evaluated using frozen sections of the SMA margin in some facilities. Nirsgke et al. reported that long-term survival was improved by re-resecting the positive surgical margin found using frozen section to achieve R0 resection [30]. However, many studies reported that intraoperative frozen section-based re-resection of R1 margins does not improve overall survival for patients with PDAC [31–33].

In patients with pancreatic head cancer, the extent of cancer infiltration varies depending on the tumor position (e.g., the difference between the pancreatic head and uncinate process) [34–37]. The development of MDCT has made it possible to confirm the accurate infiltration range of pancreatic head cancer, which can simulate the dissection range preoperatively to achieve R0 resection [38–41]. However, the extent of tumor infiltration cannot be accurately confirmed during surgery. It is often difficult to distinguish between areas of inflammatory changes and cancer infiltration during resection. Even if the "dissection to achieve R0 resection" range is simulated based on the MDCT evaluation, it is difficult to identify the range intraoperatively. Therefore, anatomical structures, such as layers, arteries, and veins, are commonly identified during surgery as anatomical landmarks to determine the dissection region [29,42].

#### **6. Anatomical Landmarks Used to Determine the Appropriate Dissection Range at Each Surgical Site**

We summarize below the anatomical structures that can be used as landmarks at each surgical site for achieving R0 resection based on preoperative diagnostic imaging.

#### *6.1. Dissection around the Hepatoduodenal Ligament and Common Hepatic Artery*

The dissection range around the hepatoduodenal mesentery and common hepatic artery (CHA) may need to be altered according to the tumor location. Uncinate process cancer invades the SMA mainly through the second part of the PLph II (equivalent to Area B) [29,34,36,43] (Figure 1B). However, in pancreatic head cancer, infiltration and lymph node metastasis around the CHA and hepatoduodenal ligament are observed [17,36]. There are 8a lymph nodes on the anterior side of the CHA, which must be removed to expose the CHA, proper hepatic artery (PHA), gastroduodenal artery (GDA), and portal vein (PV) at the superior border of the pancreas.

There is a left celiac ganglion on the right side of the root of the CHA and SMA, and nerve and fibrous tissues (NFTs) spread from the left celiac ganglion to the head of the pancreas and hepatoduodenal ligament (Area A, Figure 1C). These NFTs are divided into NFTs (PLph I) that pass through the dorsal side of the GDA (Figure 1B) and toward the upper edge of the head of the pancreas, and NFTs that pass through the dorsal side of the PHA and extend to the hepatoduodenal ligament [44] (Figure 1C). NFTs spreading to the hepatoduodenal ligament include 8p, 12p, and lymph nodes wrapped in adipose tissue [44]. These NFT regions need to be dissected when attempting complete skeletonization of the PV around the hepatic arteries around the hepatoduodenal ligament. If uncinate process cancer infiltrates around the SMA root and exposure of the CHA root is attempted, these NFT regions also need to be dissected. On the other hand, if no tumor extension is observed around the hepatoduodenal ligament and/or SMA root and CHA root, it is anatomically possible to preserve these NTF regions (Figure 2A–E).

**Figure 2.** (**A**) Tumor extension is observed near the CHA root and SMA root on the preoperative MDCT findings. (**B**) Tumor extension is observed near the hepatoduodenal ligament. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1 and 2 can be selected based on the anatomical structure, depending on tumor extension toward to the hepatoduodenal ligament, CHA root, and SMA root. (**D**) Cutting line for dissection range 1. (**E**) Cutting line for dissection range 2. MDCT: multi-detector computed tomography; CBD: common bile duct; CA: celiac artery; CHA: common hepatic artery; SMA: superior mesenteric artery; SMV: superior mesenteric vein. **Figure 2.** (**A**) Tumor extension is observed near the CHA root and SMA root on the preoperative MDCT findings. (**B**) Tumor extension is observed near the hepatoduodenal ligament. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1 and 2 can be selected based on the anatomical structure, depending on tumor extension toward to the hepatoduodenal ligament, CHA root, and SMA root. (**D**) Cutting line for dissection range 1. (**E**) Cutting line for dissection range 2. MDCT: multi-detector computed tomography; CBD: common bile duct; CA: celiac artery; CHA: common hepatic artery; SMA: superior mesenteric artery; SMV: superior mesenteric vein.

#### *6.2. Posterior Dissection 6.2. Posterior Dissection*

Few studies have described the appropriate range of posterior dissection for pancreatic head cancer. In extended lymph node dissection, including the para-aortic lymph nodes, the inferior vena cava, left renal vein, and anterior surface of the aorta are exposed. However, periaortic lymph node metastasis is now categorized as distant metastasis [15]. Prophylactic periaortic lymph node dissection is not generally performed for resectable PDAC. Delpero et al. investigated the association between each margin status and prognosis in a multicenter prospective study of 150 patients who underwent macroscopic margin-free PD. They showed that the R1 rate was 23%, while only 7% had R1 at the posterior Few studies have described the appropriate range of posterior dissection for pancreatic head cancer. In extended lymph node dissection, including the para-aortic lymph nodes, the inferior vena cava, left renal vein, and anterior surface of the aorta are exposed. However, periaortic lymph node metastasis is now categorized as distant metastasis [15]. Prophylactic periaortic lymph node dissection is not generally performed for resectable PDAC. Delpero et al. investigated the association between each margin status and prognosis in a multicenter prospective study of 150 patients who underwent macroscopic

margin-free PD. They showed that the R1 rate was 23%, while only 7% had R1 at the posterior margin; in addition, they reported that posterior R1 was not a prognostic factor [45]. Therefore, "dissection to achieve R0 resection" may not be necessary. margin; in addition, they reported that posterior R1 was not a prognostic factor [45]. Therefore, "dissection to achieve R0 resection" may not be necessary.

*Cancers* **2021**, *13*, x FOR PEER REVIEW 8 of 16

There is a fusion fascia between the posterior side of the pancreatic head and the anterior side of the vena cava and the aorta, which is called the fusion fascia of Treitz [46]. There is loose connective tissue at the anterior surface of this fusion fascia, which can be easily peeled off. If posterior infiltration is not found on the preoperative computed tomography image, this fusion fascia becomes a good anatomical landmark for indicating the range of posterior dissection. If posterior infiltration is suspected before resection and dissection with a surgical margin is needed, the anterior surface of the vena cava, renal vein, and aorta become anatomical landmarks (Figure 3A–E). There is a fusion fascia between the posterior side of the pancreatic head and the anterior side of the vena cava and the aorta, which is called the fusion fascia of Treitz [46]. There is loose connective tissue at the anterior surface of this fusion fascia, which can be easily peeled off. If posterior infiltration is not found on the preoperative computed tomography image, this fusion fascia becomes a good anatomical landmark for indicating the range of posterior dissection. If posterior infiltration is suspected before resection and dissection with a surgical margin is needed, the anterior surface of the vena cava, renal vein, and aorta become anatomical landmarks (Figure 3A–E).

**Figure 3.** (**A**) Tumor extension to the posterior side of the pancreatic head is observed on the preoperative MDCT findings. (**B**) Tumor extension to the posterior side of the pancreatic head is not observed. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1 and 2 can be selected based on the anatomical structure, depending on the range of posterior infiltration. (**D**) Surgical findings at dissection range 1. (**E**) Surgical findings at dissection range 2. MDCT: multi-detector computed tomography; VC: vena cava; LRV: left renal vein; SMA: superior mesenteric artery; SMV: superior mesenteric vein. **Figure 3.** (**A**) Tumor extension to the posterior side of the pancreatic head is observed on the preoperative MDCT findings. (**B**) Tumor extension to the posterior side of the pancreatic head is not observed. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1 and 2 can be selected based on the anatomical structure, depending on the range of posterior infiltration. (**D**) Surgical findings at dissection range 1. (**E**) Surgical findings at dissection range 2. MDCT: multi-detector computed tomography; VC: vena cava; LRV: left renal vein; SMA: superior mesenteric artery; SMV: superior mesenteric vein.

The SMA margin is the most important factor for achieving R0 resection, especially

*6.3. Dissection around the Superior Mesenteric Artery* 

#### *6.3. Dissection around the Superior Mesenteric Artery*

The SMA margin is the most important factor for achieving R0 resection, especially in uncinate process cancer, because the tumor mainly spreads behind the SMA [42,47,48]. It is difficult to understand the anatomy around the SMA during surgery because it is very complex. Recently, region between the SMA and the uncinate process has been called the "mesopancreas" [49–51]. Many surgical procedures for complete dissection of the mesopancreas have been reported [41,52–57]. However, the range of dissection varies, and the standard dissection range remains unclear. Dense connective tissues exist around the pancreatic head, which is composed of intensive nerve and fibrous tissues (NFTs). It is generally considered that cancer spreads in these areas.

The SMA is covered with NFTs called the SMA nerve plexus. The hard NFTs spread to the uncinate process from the SMA nerve plexus, which is termed as the "pancreatic head plexus II" in the Japanese Classification of Pancreatic Carcinoma [15,43]. Previously, right half-circumferential dissection of the SMA nerve plexus was performed at many facilities [54]. However, extensive dissection of the nerve plexus around the SMA often causes severe diarrhea, which may lead to delays in the induction of adjuvant chemotherapy. Jang et al. conducted a randomized clinical trial comparing extended surgery with right half-circumferential dissection of the SMA nerve plexus and standard surgery without dissection, and revealed that there was no difference in prognosis between the two groups [5]. In their study, the number 14 lymph node was not dissected in the standard group, and the dissection range around the SMA was not clearly described [58,59].

Recently, PD with complete preservation of the SMA nerve plexus has been commonly performed to avoid severe postoperative diarrhea. However, no criteria have been established to indicate an appropriate dissection range for achieving R0 resection in PD with preservation of the SMA nerve plexus. The inferior pancreaticoduodenal artery (IPDA) becomes a good anatomical landmark during the dissection around the SMA [60–63]. The IPDA forms a common trunk with the first jejunal artery in most cases (J1A) [61,64]. The dissection range can be determined during surgery based on the path of this artery. Various approaches using the IPDA, J1A, and their common arteries as landmarks have been reported for dissection around the SMA [42,56,65,66]. Inoue et al. [42] standardized the anatomical range at levels I–III, depending on the type of tumor, based on the position of the IPDA as an anatomical landmark. They reported that standardizing the dissection range reduced the operative time and blood loss in a study of 162 patients who underwent PD. Of note, the IPDA is covered with intensive NFTs and cannot be identified before initiating the SMA dissection [29,43,49]. In contrast, uncinate process cancer spreads in these intensive NFTs [17,29]. Therefore, alternative anatomical landmarks are needed for the complete dissection of these intensive NFTs in PD with preserving the SMA nerve plexus. Nagakawa et al. [29] evaluated the cancer extension of these areas using pathological specimens from 78 patients who underwent PD for resectable PDAC. According to their results, cancer invasion and/or lymph node metastasis was observed in 14.1% of NFTs (Area C) spreading to the left side of the IPDA root and in 44.9% of NFTs (Area D) spreading between the inferior side of the uncinate process and the posterior side of the jejunal mesentery (Figures 4A–E and 5A–E).

**Figure 4.** (**A**) Tumor extension to the SMA is not observed on the preoperative MDCT findings. (**B**) Tumor extension to the posterior side of the SMA is observed. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1, 2, and 3 can be selected based on the anatomical structure, depending on the range of posterior infiltration. (**D**) Cutting line for dissection range 1. (**E**) Cutting line for dissection range 3. MDCT: multi-detector computed tomography; SMA: superior mesenteric artery; IPDA: inferior pancreaticoduodenal artery; J1A: first jejunal artery; UP: uncinate process; 3rd DU: third portion of duodenum; JE: jejunum. **Figure 4.** (**A**) Tumor extension to the SMA is not observed on the preoperative MDCT findings. (**B**) Tumor extension to the posterior side of the SMA is observed. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1, 2, and 3 can be selected based on the anatomical structure, depending on the range of posterior infiltration. (**D**) Cutting line for dissection range 1. (**E**) Cutting line for dissection range 3. MDCT: multi-detector computed tomography; SMA: superior mesenteric artery; IPDA: inferior pancreaticoduodenal artery; J1A: first jejunal artery; UP: uncinate process; 3rd DU: third portion of duodenum; JE: jejunum.

**Figure 5.** (**A**) Tumor extension on the dorsal side of the jejunal mesentery is not observed on the preoperative MDCT findings. (**B**) Tumor extension on the dorsal side of the jejunal mesentery is observed. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1, 2, and 3 can be selected based on the anatomical structure, depending on the range of posterior infiltration. (**D**) Cutting line for dissection range 1. (**E**) Cutting line for dissection range 2. MDCT: multi-detector computed tomography; UP: uncinate process; SMA: superior mesenteric artery; SMV: superior mesenteric vein; PDJV: proximal dorsal jejunal vein; J2A: second jejunal artery. **Figure 5.** (**A**) Tumor extension on the dorsal side of the jejunal mesentery is not observed on the preoperative MDCT findings. (**B**) Tumor extension on the dorsal side of the jejunal mesentery is observed. (**C**) Determination of the dissection range based on the MDCT findings. Dissection ranges 1, 2, and 3 can be selected based on the anatomical structure, depending on the range of posterior infiltration. (**D**) Cutting line for dissection range 1. (**E**) Cutting line for dissection range 2. MDCT: multi-detector computed tomography; UP: uncinate process; SMA: superior mesenteric artery; SMV: superior mesenteric vein; PDJV: proximal dorsal jejunal vein; J2A: second jejunal artery.

#### *6.4. Portal Vein and/or Superior Mesenteric Vein Resection 6.4. Portal Vein and/or Superior Mesenteric Vein Resection*

PV and/or superior mesenteric vein (SMV) resection for patients with PV involvement has been generally accepted with survival benefit of pancreatic cancer [41,67–73]. The extent of PV infiltration can be diagnosed by preoperative MDCT, and the need for preoperative resection of the PV can be predicted in advance. However, there are cases in which portal vein infiltration is suspected during surgery, even if MDCT does not show tumor infiltration. In addition, it is difficult to distinguish between tumor-related fibrosis and tumor infiltration in the venous wall, and the NCCN guidelines recommend performing PV resections if tumor infiltration is suspected [74]. PV and/or superior mesenteric vein (SMV) resection for patients with PV involvement has been generally accepted with survival benefit of pancreatic cancer [41,67–73]. The extent of PV infiltration can be diagnosed by preoperative MDCT, and the need for preoperative resection of the PV can be predicted in advance. However, there are cases in which portal vein infiltration is suspected during surgery, even if MDCT does not show tumor infiltration. In addition, it is difficult to distinguish between tumor-related fibrosis and tumor infiltration in the venous wall, and the NCCN guidelines recommend performing PV resections if tumor infiltration is suspected [74].

PDAC often extends to the periphery of the SMV trunk, and the first jejunal vein (J1V) and second jejunal vein (JV) or later branches (J2, 3V) are involved with the tumor. Nevertheless, the resectability of PDAC with JV involvement remains unclear. The NCCN guidelines indicated that "unreconstructible PV/SMV due to tumor involvement or occlusion" is classified as unresectable pancreatic cancer [74]. Some surgeons choose to perform PDAC often extends to the periphery of the SMV trunk, and the first jejunal vein (J1V) and second jejunal vein (JV) or later branches (J2, 3V) are involved with the tumor. Nevertheless, the resectability of PDAC with JV involvement remains unclear. The NCCN guidelines indicated that "unreconstructible PV/SMV due to tumor involvement or occlusion" is classified as unresectable pancreatic cancer [74]. Some surgeons choose to

perform aggressive treatment such as PV/SMV resection with J1V and J2, 3V resection in patients with PDAC [75,76]. However, since the JV is thin, there is concern about the risk of complications, such as portal vein stenosis after portal vein reconstruction [77,78]. Additionally, the survival benefit of PV/SMV resection with JV resection remains unclear. Therefore, it is necessary to clarify the surgical safety and survival benefits of PV/SMV resection with JV resection (Figure 6). aggressive treatment such as PV/SMV resection with J1V and J2, 3V resection in patients with PDAC [75,76]. However, since the JV is thin, there is concern about the risk of complications, such as portal vein stenosis after portal vein reconstruction [77,78]. Additionally, the survival benefit of PV/SMV resection with JV resection remains unclear. Therefore, it is necessary to clarify the surgical safety and survival benefits of PV/SMV resection with JV resection (Figure 6).

**Figure 6.** Cutting line for portal vein, superior mesenteric vein, and jejunal vein resection. PV: portal vein; SV: splenic vein; SMV: superior mesenteric vein; JV: jejunal vein. **Figure 6.** Cutting line for portal vein, superior mesenteric vein, and jejunal vein resection. PV: portal vein; SV: splenic vein; SMV: superior mesenteric vein; JV: jejunal vein.

Several running patterns of the J1V have been reported. In 74–99% of J1Vs, the JV flows out from the dorsal side of the SMV, branches off several IPDVs along the uncinate process, passes through the dorsal side of the superior mesenteric artery, and extends to the jejunal mesentery [56,75,79,80]. It is also termed the proximal dorsal JV (PDJV) [56,75]. As the PDJV is in contact with the uncinate process, some surgeons routinely resect the PDJV without reconstruction to ensure a surgical margin, even if combined PV/SMV re-Several running patterns of the J1V have been reported. In 74–99% of J1Vs, the JV flows out from the dorsal side of the SMV, branches off several IPDVs along the uncinate process, passes through the dorsal side of the superior mesenteric artery, and extends to the jejunal mesentery [56,75,79,80]. It is also termed the proximal dorsal JV (PDJV) [56,75]. As the PDJV is in contact with the uncinate process, some surgeons routinely resect the PDJV without reconstruction to ensure a surgical margin, even if combined PV/SMV resection is not required [75,76] (Figure 5).

#### **7. Conclusions**

section is not required [75,76] (Figure 5).

**7. Conclusions**  The role of surgery has changed dramatically in the current treatment, where multimodal treatment has become important to improve the prognosis of resectable PDAC. Now that effective preoperative and postoperative chemotherapy has been established, it may be necessary to reconsider the areas treated with chemotherapy and the areas treated with surgery. On the other hand, many studies have described that R0 resection is needed even in patients receiving adjuvant therapy. The appropriate dissection range for R0 resection can be simulated preoperatively with MDCT imaging. Therefore, surgeons need to perform a more accurate dissection, balancing both R0 resection and the introduction The role of surgery has changed dramatically in the current treatment, where multimodal treatment has become important to improve the prognosis of resectable PDAC. Now that effective preoperative and postoperative chemotherapy has been established, it may be necessary to reconsider the areas treated with chemotherapy and the areas treated with surgery. On the other hand, many studies have described that R0 resection is needed even in patients receiving adjuvant therapy. The appropriate dissection range for R0 resection can be simulated preoperatively with MDCT imaging. Therefore, surgeons need to perform a more accurate dissection, balancing both R0 resection and the introduction of adjuvant therapy, based on the precise anatomy.

of adjuvant therapy, based on the precise anatomy. **Author Contributions:** Details on the design of this review were discussed with Y.N.; N.N. (Naoya Nakagawa); C.T.; S.K.; H.O.; K.S.; N.N. (Nobuhiko Nakagawa); Y.H.; T.S. and M.H. The literature was searched and reviewed by Y.N.; N.N. (Naoya Nakagawa), and C.T. The draft of the manuscript **Author Contributions:** Details on the design of this review were discussed with Y.N.; N.N. (Naoya Nakagawa); C.T.; S.K.; H.O.; K.S.; N.N. (Nobuhiko Nakagawa); Y.H.; T.S. and M.H. The literature was searched and reviewed by Y.N.; N.N. (Naoya Nakagawa), and C.T. The draft of the manuscript was critiqued by Y.N.; I.U. and T.Y. This review was validated by K.K. and A.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Optimal Preoperative Multidisciplinary Treatment in Borderline Resectable Pancreatic Cancer**

**Nana Kimura 1,†, Suguru Yamada 2,† , Hideki Takami <sup>2</sup> , Kenta Murotani <sup>3</sup> , Isaku Yoshioka <sup>1</sup> , Kazuto Shibuya <sup>1</sup> , Fuminori Sonohara <sup>2</sup> , Yui Hoshino <sup>1</sup> , Katsuhisa Hirano <sup>1</sup> , Toru Watanabe <sup>1</sup> , Hayato Baba <sup>1</sup> , Kosuke Mori <sup>1</sup> , Takeshi Miwa <sup>1</sup> , Mitsuro Kanda <sup>2</sup> , Masamichi Hayashi <sup>2</sup> , Koshi Matsui <sup>1</sup> , Tomoyuki Okumura <sup>1</sup> , Yasuhiro Kodera <sup>2</sup> and Tsutomu Fujii 1,\***


**Simple Summary:** For borderline pancreatic cancer, upfront surgery was standard in the past, and the usefulness of neoadjuvant treatment has been reported in recent years. However, few studies have been conducted to date on whether there is a difference in optimal treatment between borderline resectable pancreatic cancer invading the portal vein (BR-PV) or abutting major arteries (BR-A). The objective of this study was to investigate the optimal neoadjuvant therapy for BR-PV or BR-A. We retrospectively analyzed 88 patients with BR-PV and 111 patients with BR-A. In this study, we found that neoadjuvant treatment using new chemotherapy (FOLFIRINOX or gemcitabine along with nab-paclitaxel) is essential for improving the prognosis of BR pancreatic cancer. These findings suggest that prognosis may be prolonged by maintaining good nutritional status during preoperative treatment.

**Abstract:** *Background:* The objective of this study was to investigate the optimal neoadjuvant therapy (NAT) for borderline resectable pancreatic cancer invading the portal vein (BR-PV) or abutting major arteries (BR-A). *Methods:* We retrospectively analyzed 88 patients with BR-PV and 111 patients with BR-A. *Results:* In BR-PV patients who underwent upfront surgery (*n* = 46)/NAT (*n* = 42), survival was significantly better in the NAT group (3-year overall survival (OS): 5.8%/35.5%, *p* = 0.004). In BR-A patients who underwent upfront surgery (*n* = 48)/NAT (*n* = 63), survival was also significantly better in the NAT group (3-year OS:15.5%/41.7%, *p* < 0.001). The prognosis tended to be better in patients who received newer chemotherapeutic regimens, such as FOLFIRINOX and gemcitabine with nab-paclitaxel. In 36 BR-PV patients who underwent surgery after NAT, univariate analysis revealed that normalization of tumor marker (TM) levels (*p* = 0.028) and preoperative high prognostic nutritional index (PNI) (*p* = 0.022) were significantly associated with a favorable prognosis. In 39 BR-A patients who underwent surgery after NAT, multivariate analysis revealed that preoperative PNI > 42.5 was an independent prognostic factor (HR: 0.15, *p* = 0.014). *Conclusions:* NAT using newer chemotherapy is essential for improving the prognosis of BR pancreatic cancer. These findings suggest that prognosis may be prolonged by maintaining good nutritional status during preoperative treatment.

**Citation:** Kimura, N.; Yamada, S.; Takami, H.; Murotani, K.; Yoshioka, I.; Shibuya, K.; Sonohara, F.; Hoshino, Y.; Hirano, K.; Watanabe, T.; Baba, H.; et al. Optimal Preoperative Multidisciplinary Treatment in Borderline Resectable Pancreatic Cancer. *Cancers* **2021**, *13*, 36. https://dx.doi.org/10.3390/cancers 13010036

Received: 14 December 2020 Accepted: 21 December 2020 Published: 24 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/).

**Keywords:** pancreatic cancer; borderline resectable; neoadjuvant treatment; chemoradiotherapy; prognostic nutritional index

#### **1. Introduction**

Despite considerable improvements in diagnostic and therapeutic options, pancreatic ductal adenocarcinoma (PDAC) mostly remains a fatal disease worldwide [1]. Radical resection without residual tumor remains the only established curative treatment for PDAC. However, much more intervention is required beyond resection alone. A simple explanation for the poor consequences after resection is that almost every patient has microscopic disease remaining [2]. When a patient is diagnosed with PDAC and the optimal treatment strategy is considered, it is common to make a decision based on the resectable classification rather than the stage classification. The National Comprehensive Cancer Network (NCCN), an alliance of 25 cancer centers in the United States, have proposed a resectable classification for pancreatic cancer [3]. However, the NCCN guidelines are revised and updated annually and are considered to be very complex; thus, utilizing the NCCN classification system for resectability in clinical practice is difficult. Therefore, the Japan Pancreas Society (JPS) proposed novel and simplified resectability criteria in 2016 [4] based on the most recent NCCN guidelines [3].

The JPS published the 7th edition of the Classification of Pancreatic Carcinoma, and a unique resectable classification for borderline resectable (BR) was proposed (BR-A: BR-PDAC due to the infiltration of celiac and/or superior mesentery arteries, BR-PV: due only to the infiltration of the portal system). BR pancreatic cancer is a distinct subset of locally advanced pancreatic cancer first identified by Varadhachary et al. in 2006 [5]. It was hoped that the BR group would represent a subset of pancreatic cancer whose outcomes might be intermediate between the outcomes of patients with radiologically and technically resectable (R) and unresectable (UR) disease. With currently available operative techniques, patients with BR cancer are at high risk for margin-positive resection [6]. Therefore, the criteria for resectability are clinically important for determining the need for preoperative (neoadjuvant) systemic therapy and/or local-regional chemoradiation to maximize the potential for R0 resection and to avoid R2 resection [7]. For BR pancreatic cancer, upfront surgery was standard in the past, and the usefulness of neoadjuvant treatment (NAT) has been reported in recent years [8–10]. However, few studies have been conducted to date on whether there is a difference in optimal treatment between BR-PV and BR-A.

The objective of this study was to investigate the optimal preoperative multidisciplinary treatment for BR pancreatic cancer. Patients who had received treatment for BR-PDAC at two regional high-volume centers were reviewed retrospectively, and we analyzed survival differences among subgroups defined based on this novel classification system of resectability.

#### **2. Results**

#### *2.1. Cohort Outline*

We identified 199 patients who were diagnosed with BR-PDAC (Figure 1). Among them, 88 patients were diagnosed with BR-PV PDAC, and 111 patients were diagnosed with BR-A PDAC.

Of 88 BR-PV patients, 46 patients underwent upfront surgery, and 36 patients underwent resection after NAT. The other 6 patients did not undergo surgery because of chemotherapeutic failure or best supportive care. Of 111 BR-A patients, 48 patients underwent upfront surgery, and 39 patients underwent resection after NAT. The other 24 patients did not undergo surgery because of chemotherapeutic failure or best supportive care.

**Figure 1.** Study profiles and clinical courses of the enrolled patients. BR, borderline resectable; PDAC, pancreatic ductal adenocarcinoma; NAT, neoadjuvant treatment.

#### *2.2. The Clinical Characteristics of BR-PDAC Patients*

For patients who were enrolled in this study, detailed cohort demographics are summarized in Table 1. The median age was 66 years in BR-PV patients and 67 years in BR-A patients. Preoperative image examination revealed that the location of the tumor was dominant (BR-PV: 95%, BR-A: 75%) on the head side in both BR-PV and BR-A; thus, pancreatic head resection tended to be more frequent (BR-PV: 84%, BR-A: 60%).

In the BR-PV patients, 26 (30%) patients were treated with newer chemotherapeutic regimens such as FOLFIRINOX (FFX) and gemcitabine along with nab-paclitaxel (GnP). The median length of therapy was 2.1 months. In the BR-A patients, 36 (32%) patients were treated with newer chemotherapeutic regimens with a median length of 2.7 months.

The median baseline CA19-9 level at diagnosis was higher than the median at surgery. Additionally, both the BR-PV and BR-A groups had lower median CA19-9 levels at operation in patients who underwent surgery after NAT than in those who underwent upfront surgery. In both BR-PV and BR-A patients, approximately 20% of patients had a ≥90% decrease in CA19-9 levels compared to that before NAT. This suggests that preoperative NAT may be expected to significantly reduce tumor markers (TMs), as in previous reports [11].

The median baseline nutritional parameters at operation were as follows (in BR-PV/BR-A): controlling nutritional status (CONUT): 2/2, Glasgow prognostic score (GPS): 0/0, modified GPS (mGPS): 0/0, neutrophil/lymphocyte ratio (NLR): 2.4/2.5, platelet/ lymphocyte ratio (PLR): 129.2/83.0, prognostic nutritional index (PNI): 46.0/44.5, lymphocyte/monocyte ratio (LMR): 3.6/3.8, systemic immune inflammation index (SII): 380.1/482.8, and C-reactive protein (CRP)/albumin ratio: 0.07/0.03.

For patients with BR-PDAC who were underwent surgery, detailed cohort demographics are summarized in Table 2. The median age was 65 years in BR-PV patients and 67 years in BR-A patients. In operation, venous resection was performed in 72 (88%) patients with BR-PV and 62 (71%) patients with BR-A. Moreover, arterial resection was performed in 5 (6%) patients with BR-PV and 12 (14%) patients with BR-A.

In addition, we compared the backgrounds of patients who underwent upfront surgery and those who underwent NAT. Details are shown in Table 3. For both BR-A and BR-PV, the CA19-9 level at operation was lower in the NAT group. There was no significant difference in preoperative nutritional status.


**Table 1.** Baseline characteristics of patients with BR-PV and BR-A.

\* values are median (range). CA19-9, carbohydrate antigen 19-9; FFX, FOLFIRINOX; GnP, gemcitabine along with nab-paclitaxel; GS, gemcitabine along with S-1; CONUT, controlling nutritional status; GPS, Glasgow prognostic score; mGPS, modified Glasgow prognostic score; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio; PNI, prognostic nutritional index; LMR, lymphocyte/monocyte ratio; SII, systemic immune inflammation index; CRP, C-reactive protein; Alb, albumin; PRBC, packed red blood cells.

**Table 2.** Baseline characteristics of patients with BR-PDAC who underwent resection.



**Table 2.** *Cont*.

\* values are median (range). CA19-9, carbohydrate antigen 19-9; FFX, FOLFIRINOX; GnP, gemcitabine along with nab-paclitaxel; GS, gemcitabine along with S-1; PRBC, packed red blood cells.

**Table 3.** Baseline characteristics of patients with BR-PDAC who underwent upfront surgery or NAT.


\* Values are median (range). CA19-9, carbohydrate antigen 19-9; CONUT, controlling nutritional status; N/A, not available; GPS, Glasgow prognostic score; mGPS, modified Glasgow prognostic score; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio; PNI, prognostic nutritional index; LMR, lymphocyte/monocyte ratio; SII, systemic immune inflammation index; CRP, C-reactive protein; Alb, albumin.

#### *2.3. Comparison of Prognosis of Upfront Surgery vs. Neoadjuvant Treatment by Intention to Treat Analysis*

In BR-PV patients who underwent upfront surgery (*n* = 46)/NAT (*n* = 42), survival was significantly better in the NAT group (*p* = 0.004) (Figure 2). In BR-A patients who underwent upfront surgery (*n* = 48)/NAT (*n* = 63), survival was significantly better in the NAT group (*p* < 0.001). This analysis was performed by intention-to-treat analysis.

**Figure 2.** The overall survival in comparison between patients treated with and without NAT in the (**a**) BR-PV and (**b**) BR-A groups. The prognosis of patients treated with NAT was significantly better than that of patients treated without NAT in the BR-PV and BR-A groups (*p* = 0.004 and *p* < 0.001). NAT, neoadjuvant treatment; UFS, upfront surgery; MST, median survival time; HR, hazard ratio; CI, confidence interval.

The 36-month (3-year) OS rates with upfront surgery and NAT were 5.8% versus 35.5% in BR-PV patients and 15.5% versus 41.7% in BR-A patients, respectively.

#### *2.4. Comparison of Regimens in Neoadjuvant Treatment Induction Cases*

We compared the regimens of neoadjuvant treatment in each group (Table 4).

In BR-PV patients who underwent FFX/GnP (*n* = 26) vs. gemcitabine (GEM)/S-1 (*n* = 2) vs. GEM/S-1 with radiotherapy (RT) (*n* = 14), the median survival times (MSTs) were 32.9, 10.0 and 20.6 months, respectively, and the prognosis tended to be better in the FFX/GnP group. The number of resected cases was 36 (86%).

In BR-A patients who underwent FFX/GnP (*n* = 29) vs. FFX/GnP with RT (*n* = 7) vs. GEM/S-1 (*n* = 10) vs. GEM/S-1 with RT (*n* = 17), the MSTs were 35.4, 18.7, 43.2 and 19.7 months, respectively, with a better prognosis in the FFX/GnP group. The number of resected cases was 39 (62%).

The R0 rate tended to be higher in regimens with RT.


**Table 4.** Comparison of regimens in patients who underwent NAT.

Old NAC means neoadjuvant chemotherapy including gemcitabine, S-1, and GEM with S-1; FFX, FOLFIRINOX; GnP, gemcitabine along with nab-paclitaxel; RT, radiotherapy; NAC, neoadjuvant chemotherapy; MST, median survival time.

#### *2.5. Prognostic Factors in Patients Who Underwent Resection after NAT*

#### 2.5.1. Definition of Cutoff Values for PNI

Receiver operating characteristic (ROC) curve analysis was performed with data from 36 BR-PV PDAC patients who underwent surgical resection between January 2002 and December 2018 to examine the association between PNI and 2-year survival. The area under the curve (AUC) was 0.728, and the best cutoff value was calculated as 42.65 (Figure 3a). Moreover, ROC curve analysis was performed with data from 39 BR-A PDAC patients. The AUC curve was 0.820, and the best cutoff value was calculated as 42.50 (Figure 3b). We eventually determined that the cutoff value for PNI was 42.5.

**Figure 3.** ROC analysis for the prediction of 2-year survival according to the preoperative PNI. (**a**) The AUC was 0.820 in BR-PV patients. (**b**) The AUC was 0.728 in BR-A patients. AUC, area under the curve; PNI, prognostic nutritional index.

2.5.2. Univariate and Multivariate Analyses of Prognostic Factors in BR-PDAC Patients Who Underwent Resection after Neoadjuvant Treatment

Table 5 shows the results of univariate analysis of prognostic factors in BR-PDAC patients who underwent resection after NAT. The cutoff values for continuous variables except preoperative PNI were determined using median values of all BR-PDAC patients who underwent resection after NAT.

In 36 BR-PV patients who underwent surgery after NAT, univariate analysis of overall survival revealed that normalization of TM levels (*p* = 0.028), preoperative GPS = 0 (*p* = 0.025), and preoperative high PNI (*p* = 0.022) were significantly associated with better prognosis. There was no significant difference in the multivariate analysis.

In 39 BR-A patients who underwent surgery after NAT, univariate analysis revealed that normalization of TM levels (*p* = 0.033), preoperative high PNI (*p* = 0.013), and intraoperative blood loss ≤ 830 mL (*p* = 0.013) were significantly associated with better prognosis. Multivariate analysis showed that preoperative PNI > 42.5 was an independent prognostic factor (HR: 0.15, *p* = 0.014). There was no correlation between the length of NAT and additional RT in survival in either BR-PV or BR-A.


**Table 5.** Univariate and multivariate analyses of the clinical features of BR-PDAC patients who underwent resection after NAT.


**Table 5.** *Cont*.

CA19-9, carbohydrate antigen 19-9; NAT, neoadjuvant treatment; Alb, albumin; CONUT, controlling nutritional status; GPS, Glasgow prognostic score; mGPS, modified Glasgow prognostic score; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio; PNI, prognostic nutritional index; LMR, lymphocyte/monocyte ratio; SII, systemic immune inflammation index; CRP, C-reactive protein; Alb, albumin; \* *p* < 0.05.

> 2.5.3. Prognosis of BR-PDAC Patients Who Underwent Resection after Neoadjuvant Treatment Based on PNI

In BR-PV patients who underwent resection after NAT (*n* = 36), survival was significantly better in the high PNI (preoperative PNI > 42.50) group (*p* = 0.029, HR:0.16, 95%CI:0.03–0.83) (Figure 4a). In BR-A patients who underwent resection after NAT (*n* = 39), survival was significantly better in the high PNI (preoperative PNI > 42.50) group (*p* = 0.012, HR:0.13, 95%CI:0.03–0.64) (Figure 4b).

Moreover, comparing the high preoperative PNI and low preoperative PNI, there was no statistically significant difference regarding postoperative complications (Clavien–Dindo grade III or more) in both BR-PV and BR-A patients (*p* = 0.644 and *p* = 0.580, respectively).

**Figure 4.** Comparison of the overall survival between high preoperative PNI and low preoperative PNI in the (**a**) BR-PV and (**b**) BR-A groups. The prognosis of patients with high preoperative PNI was significantly better than that of patients with low preoperative PNI in the BR-PV and BR-A patients (*p* = 0.029 and *p* = 0.012). PNI, prognostic nutritional index.

#### **3. Discussion**

There have been many analytical studies on R-PDAC and UR-PDAC, but few have focused on BR-PDAC. The usefulness of NAT for BR-PDAC has been highlighted in several articles [8–11]. Unfortunately, previous reports often analyzed mixed cohorts of patients, including those with BR and locally advanced UR-PDAC, those with BR-PDAC due to the infiltration of celiac and/or superior mesentery arteries (BR-A) and those with only infiltration of the portal system (BR-PV) [12,13]. The surgical strategy and outcome definitely differ between PDAC abutted to the major arteries and PDAC exclusively involving the PV system [14]. Murakami et al. reported that the BR-PV group had a significantly more favorable overall survival than the BR-A group in an analysis of BR patients who underwent upfront surgery [15]. Thus, it seems inappropriate to discuss the efficacy of the treatment strategy using such admixture.

In the present study, we differentiated between BR-A and BR-PV and analyzed the optimal preoperative multidisciplinary treatment and nutritional status before and after NAT for each type. There have been no comprehensive analyses focusing on surgical strategy for this cohort.

#### *3.1. BR-PV*

We retrospectively reviewed 88 patients with BR-PV PDAC. The results showed that the prognosis of BR-PV patients who underwent resection after NAT was significantly better than that of patients who underwent upfront surgery without NAT.

Fujii et al. reported that neoadjuvant chemoradiotherapy (NACRT) with S-1 rather than upfront surgery improves R0 rates and increases the survival of patients with BR-PV adenocarcinoma of the pancreatic head but not that of patients with R-PDAC [6]. However, the prognosis tended to be better in the FFX/GnP group than in the NACRT with old chemotherapy group in the present study. Although only 14 BR-PV patients underwent NACRT in this study, chemotherapeutic regimens such as FFX/GnP are expected to be a promising option.

#### *3.2. BR-A*

We retrospectively reviewed 111 patients with BR-A PDAC. Similar to that of BR-PV patients, the prognosis of BR-A patients who underwent resection after NAT was significantly better than that of those who underwent upfront surgery without NAT. Moreover, in patients with BR-A, the use of NAT with FFX/GnP significantly prolonged the prognosis.

Nagakawa et al. reported that NACRT, which combines chemotherapy with GEM/S-1, with intensity modified radiotherapy (IMRT) had fewer adverse events and improved the prognosis of BR-A [16]. In addition, they also reported that the R0 resection rate after NACRT was 94.7%. In the present study, the R0 resection rate also tended to be higher in patients who underwent additional RT, although additional RT failed to contribute to patient survival.

Hackert et al. reported that resection rates following FFX were 61% compared with 46% after GEM and RT in patients with locally advanced PDAC [17]. This study did not investigate NACRT, which combines new chemotherapy and RT; thus, it cannot be affirmed. However, the combination of radiation with more effective chemotherapy, such as FFX or GnP, is expected to improve the surgical consequences of BR-A patients. On the other hand, effective chemotherapy may lead to more adverse events. There is a report that NAT with FFX followed by IMRT concurrent with fixed-dose-rate GEM in BR-PDAC is feasible and tolerated [18]. Therefore, the IMRT technique may enable the application of NACRT in combination with more effective chemotherapy.

Over the past decade, newer chemotherapeutic regimens, including FFX and GnP, have emerged as new standard therapies for PDAC, which was formerly a lethal disease, and many studies have demonstrated promising survival rates [8–10,12,13]. However, there are few prospective randomized controlled studies to confirm the efficacy of NAT for patients with BR-PV and BR-A [16]. Evaluation of NAT is required for patients with BR-PV and BR-A in the setting of prospective trials.

#### *3.3. BR-PV and BR-A*

From previous reports as well as the results of this study, surgery after NAT is arguably more beneficial than upfront surgery in patients with BR-PDAC; therefore, we focused only on the patients who underwent NAT in the analysis after Section 2.4.

We demonstrated that the long-term survival of patients who underwent resection after NAT was significantly associated with good nutritional status, such as a PNI of more than 42.5 at the time of operation but not at diagnosis. Several studies have reported that preoperative nutrition indices, such as CONUT, mGPS, and PNI, are linked to the prognosis of various malignancies [19–21]. In pancreatic cancer, some indices have also been reported to have an independent association with survival in patients with resectable or BR-PDAC after pancreatectomy [22,23]. Moreover, there was a report that NAT for PDAC could aggravate nutritional status and hamper its postoperative recovery and that malnutrition might decrease the tolerance of NAT [24]. While definitive conclusions cannot be drawn from this retrospective study, these results strongly suggest the need for nutritional care during NAT in patients with PDAC. Systemic chemotherapy generally tends to worsen the patient's nutritional status as a side effect, including loss of appetite or dysgeusia [25]. Active nutritional care during NAC may minimize malnutrition, possibly improving the survival of BR-PDAC patients.

Furthermore, both the BR-PV and BR-A groups had lower median CA19-9 levels at operation in patients who underwent surgery after NAT than those who underwent upfront surgery. Although no significant difference was found in the multivariate analysis, normalization of TM levels was significantly associated with better prognosis in both BR-PV and BR-A patients. Chen et al. reported that long-term (approximately 6 months) chemotherapy after preoperative chemoradiotherapy may improve the prognosis in patients with potentially resectable/BR/UR-PDAC [26]. Satoi et al. also reported that the prognosis was prolonged by giving chemotherapy for 240 days or more in patients with UR-PDAC [27]. Conversely, the results of this study showed that the duration of NAT did not correlate with prognosis for BR-PDAC. It is suggested that the prognosis may be prolonged by surgery after the TM level is greatly reduced, not the length of NAT. Some patients with BR-PDAC became unresectable due to the progression of disease, such as distant metastasis during NAT. Therefore, assessment of TMs may be a more sensitive measure of the indications for resection than the length of treatment. Further exploration will be required for the optimal NAT duration and timing of surgery.

There are several limitations to our study. First, this study was retrospective in design with a relatively small number of patients. Second, the distribution of patients in the different treatment arms was unbalanced. Third, the number of patients receiving NAT has increased since around 2010. The proportion of patients receiving NAT and upfront surgery has changed significantly between 2002 and 2018. Fourth, indications for CRT were biased because CRT was recommended at the physicians' discretion. Finally, we have not started nutritional support for patients with NAT. We plan to explore the effect of nutritional support during NAT for patients with BR-PDAC to confirm the clinical relevance of this study. Further studies with more patients and longer observation periods are needed to evaluate the optimal and detailed strategy of multidisciplinary treatment for BR-PDAC.

In conclusion, these findings suggest that NAT followed by surgery rather than upfront surgery offers clinical benefits to patients with BR-A PDAC. Moreover, nutritional management during NAT may lead to a better prognosis.

#### **4. Materials and Methods**

#### *4.1. Study Design*

A prospectively maintained pancreatic resection database at two regional high-volume centers, Toyama University Hospital (Toyama, Japan) and Nagoya University Hospital (Nagoya, Japan), was queried to identify patients with BR PDAC who started the initial treatment between January 2002 and December 2018. This study conforms to the ethical guidelines of the World Medical Association Declaration of Helsinki-Ethical Principles for Medical Research Involving Human Subject. Written informed consent for inclusion in the study, as required by the institutional review board of both institutions, was obtained from all patients.

We retrospectively examined 199 patients with BR pancreatic cancer (88 patients with BR-PV and 111 patients with BR-A). For BR-PV and BR-A, the following points were investigated:


#### *4.2. Definitions of BR-PV PDAC and BR-A PDAC Patients*

The preoperative resectability status was categorized into R (resectable), BR-PV, BR-A, UR (unresectable)-LA, and UR-M (metastatic) according to the 7th edition of the JPS classification (Table 6) [4].


**Table 6.** Resectability criteria proposed by the JPS.

JPS, Japan Pancreas Society; SMV, superior mesenteric vein; PV, portal vein; CA, celiac artery; SMA, superior mesenteric artery; CHA, common hepatic artery; PHA, proper hepatic artery.

Patient eligibility was rigorously defined using thin-slice multidetector-row computed tomography. All images were reviewed by two or more experienced radiologists to reaffirm the preoperative staging. Consequently, 199 patients with BR-PDAC (88 patients with BR-PV and 111 patients with BR-A) were enrolled in this study.

#### *4.3. Neoadjuvant Treatment*

Some patients received NAT from diagnosis with the following regimens: GnP, FFX, modified FFX (mFFX), or GEM with oral S-1 (the oral 5-fluorouracil prodrug tegafur with oteracil and gimeracil). These chemotherapeutic regimens were selected depending on the patient's background and the period of enrollment. We performed NAT in 105 patients (42 patients with BR-PV and 63 patients with BR-A) from which informed consent was obtained depending on their condition and tumor status. CRT consisted of a photon/proton external beam with 50.4 Gy delivered in 28 fractions combined with systemic chemotherapy involving oral S-1, which was administered twice daily (80 mg/m2/day) from days 1 to 14 and from days 22 to 35.

#### *4.4. Postoperative Adjuvant Therapy*

Postoperative adjuvant chemotherapy was applied unless contraindicated by the patient's condition. In short, the patients received GEM or S-1 for 6 months according to the protocol that was available at the time of treatment [28,29]. GEM at a dose of 1000 mg/m<sup>2</sup> was administered weekly for 3 weeks followed by 1 week of rest; oral S-1 (80 mg/m2/day) was administered from days 1 to 28 followed by a 2-week rest period. Chemotherapy was initiated at <2 months after the operation in all patients who were considered eligible for the treatment. Computed tomography was routinely performed every 6 months as a postoperative follow-up imaging examination, and a blood test, including evaluation of TMs, was performed every 2 months to evaluate the recurrent disease.

#### *4.5. Data Collection*

We collected patient data from the medical records. Pretreatment factors included age, sex, body mass index, tumor size, and blood test results, including serum CA19-9 level. Preoperative factors included chemotherapeutic regimen, length of NAT, and change in CA19-9 level. Perioperative factors included surgical procedures, region of tumor, operative time, blood loss volume, blood transfusion, incidence of postoperative complications according to the Clavien–Dindo classification [30], length of hospital stay, and 90-day mortality.

The tumor-node-metastasis staging system for pancreatic tumors of the seventh edition of the Union for International Cancer Control was applied [31]. The pathological data collected included tumor grade, number of positive lymph nodes, resection margins, perineural invasion, PV invasion, and artery invasion. The surgical margin in this study denoted either the stump of the pancreas or the bile duct or the dissected plane around the pancreas as described by Staley et al. [32]. If viable cancer cells were detected microscopically at the tip of any of these sites, the surgical margin was noted as positive. If the tumor was located at a distance of >1 mm from the surgical margin, the margin was noted as negative.

#### *4.6. Nutritional Status*

In the current study, we also investigated several nutritional parameters at diagnosis and at operation, such as the GPS [33], mGPS [33], CONUT [19], PNI [22,33], NLR [33,34], PLR [33], LMR [34], and SII [35], to verify their impact on the operative outcome and the prognosis.

#### *4.7. Statistical Analysis*

A biostatistician (K.M.) was responsible for the statistical analysis. The Kaplan–Meier method was used to calculate survival rates, and the difference in survival curves was analyzed by the log-rank test. To detect prognostic factors for survival, we performed Cox proportional hazard analysis, and hazard ratios and 95% confidence intervals (CIs) were calculated. Goodness-of-fit for preoperative PNI was assessed by calculating the AUC of the ROC curve, and the optimal cutoff value was determined using the Youden index. Other cutoff values in Table 3 used their median values. Differences in nominal data

between the two groups were examined using the chi-square test or Fisher's exact test when the expected value was <5. Differences in quantitative variables were evaluated using Student's *t*-test or the Mann–Whitney *U* test if the distribution was abnormal. A *p* value < 0.05 was considered statistically significant. All statistical analyses were performed using JMP statistical software (version 14.2; SAS Institute, Cary, NC, USA).

#### **5. Conclusions**

NAT using chemotherapy such as FFX or GnP is essential for improving the prognosis of BR pancreatic cancer. This suggests that prognosis may be improved by maintaining good nutritional status during preoperative treatment, not by the length of preoperative treatment. In addition, normalization of TMs by preoperative treatment contributes to the prolongation of survival.

**Author Contributions:** Conceptualization, N.K., S.Y. and T.F.; methodology, H.T., I.Y. and K.S.; formal analysis, K.M.; investigation F.S., Y.H., K.H., T.W. and K.M. (Kosuke Mori); data curation, H.B., T.M., M.K., M.H., K.M. (Koshi Matsui) and T.O.; writing—original draft preparation, N.K., Y.H. and H.B.; writing—review and editing, N.K., S.Y. and T.F.; supervision, Y.K. and T.F.; project administration, Y.K.; funding acquisition, T.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by JSPS KAKENHI, grant number 18H02878.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Ethics Committee of University of Toyama (protocol code: R2019141 and date of approval: 2 December 2019).

**Informed Consent Statement:** Patient consent was not required based on the use of anonymized data.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Review* **Surgical Treatment of Pancreatic Ductal Adenocarcinoma**

**Kongyuan Wei and Thilo Hackert \***

Department of General, Visceral and Transplantation Surgery, University of Heidelberg, Im Neuenheimer Feld 420, 69120 Heidelberg, Germany; jasonwky@163.com **\*** Correspondence: thilo\_hackert@med.uni-heidelberg.de; Tel.: +49-6221-565150

**Simple Summary:** Surgery is the only potential cure for pancreatic ductal adenocarcinoma and should always be combined with adjuvant chemotherapy or other multimodal treatment. Besides the advances in such multimodal approaches, there has been substantial progress in surgical techniques to especially address advanced resections. These techniques include specific operative steps, such as 'artery first' or 'uncinate first' approaches as well as techniques that allow safe vascular resection and reconstruction to achieve radical tumor removal. Most recently, also minimally-invasive and robotic approaches have been adopted for pancreatic cancer surgery; however, there is no high-level evidence on these evolving techniques especially with regards to long-term results compared to conventional surgical techniques.

**Abstract:** Pancreatic ductal adenocarcinoma (PDAC) represents an aggressive tumor of the digestive system with still low five-year survival of less than 10%. Although there are improvements for multimodal therapy of PDAC, surgery still remains the effective way to treat the disease. Combined with adjuvant and/or neoadjuvant treatment, pancreatic surgery is able to enhance the five-year survival up to around 20%. However, pancreatic resection is always associated with a high risk of complications and regarded as one of the most complex fields in abdominal surgery. This review gives a summary on the surgical treatment for PDAC based on the current literature with a special focus on resection techniques.

**Keywords:** pancreatic ductal adenocarcinoma; surgical treatment; technical advances

#### **1. Introduction**

Pancreatic ductal adenocarcinoma (PDAC) still remains a big therapeutic challenge for its poor prognosis and will likely becomes the second cause of cancer death within the next decade [1,2]. Although there are mounts of advanced treatments including adjuvant chemotherapy, surgical therapy is always regarded as the most effective one to attain the long-term survival for the patients with PDAC [3]. Unfortunately, less than 20% of patients with pancreatic cancer are considered as the surgically resectable cases until now [2]. Additionally, most of the patients with metastatic disease are not suitable for resection according to the safety and efficacy affected by the historical concerns [4]. However, owing to the development of systematic chemotherapy and improvement of surgery, extended indications of the pancreatic resection are applied in clinical practice, including technical advancements as well as patient criteria such as advanced age [5].

#### *1.1. Definition of Resectability*

There are various classifications reported for the differentiation of resectable, borderlineresectable, and unresectable pancreatic cancers [6–11]. The definition of resectability is made mainly based on scientific associations as well as the MD Anderson Classification [8,10]. The AHPBA/SSO/SSAT Classification was modified by the National Comprehensive Cancer Network (NCCN) further as well as the International Study Group of Pancreatic Surgery (ISGPS) [11,12]. Therefore, resectability now is classified by the invasion

**Citation:** Wei, K.; Hackert, T. Surgical Treatment of Pancreatic Ductal Adenocarcinoma. *Cancers* **2021**, *13*, 1971. https://doi.org/ 10.3390/cancers13081971

Academic Editor: Tsutomu Fujii

Received: 17 March 2021 Accepted: 13 April 2021 Published: 20 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

cally possible.

of important adjacent vessels, especially referring to the celiac trunk, superior mesenteric artery (SMA), and the portal (PV) or superior mesenteric vein (SMV). Pancreatic cancer is regarded as resectable if there are no major vessels involved. Borderline resectable pancreatic cancer is defined as a pancreatic cancer with involvement of the portal vein and/or superior mesenteric vein and the involved segments of vessels allow resection and reconstruction (Figure 1). portant adjacent vessels, especially referring to the celiac trunk, superior mesenteric artery (SMA), and the portal (PV) or superior mesenteric vein (SMV). Pancreatic cancer is regarded as resectable if there are no major vessels involved. Borderline resectable pancreatic cancer is defined as a pancreatic cancer with involvement of the portal vein and/or superior mesenteric vein and the involved segments of vessels allow resection and reconstruction (Figure 1).

Surgery (ISGPS) [11,12]. Therefore, resectability now is classified by the invasion of im-

*Cancers* **2021**, *13*, x FOR PEER REVIEW 2 of 16

**Figure 1.** Anatomical borderline resectability, contrast enhanced CT scan, and coronary reformatting. Pancreatic head cancer with contact to superior mesenteric vein/portal vein confluence (white circle), vascular reconstruction techni-**Figure 1.** Anatomical borderline resectability, contrast enhanced CT scan, and coronary reformatting. Pancreatic head cancer with contact to superior mesenteric vein/portal vein confluence (white circle), vascular reconstruction technically possible.

Furthermore, if the superior mesenteric artery or the celiac trunk are invaded, pancreatic tumors are considered as locally advanced and unresectable, however arterial resections and reconstructions can be performed by experienced surgeons. Actually, for determination of resectability, the relationship between the tumor and mesenteric/hepatic vessels is the critical topic to obtain R0 resection [13–15]. Hence, through preoperative staging and imaging, pancreatic tumors are divided into four types: resectable, borderline resectable, locally advanced and metastatic. Currently, upfront surgery is recommended in resectable pancreatic cancer [7,11,12]. In contrary to the surgical and anatomical considerations to evaluate resectability, the International Association of Pancreatology developed a more comprehensive definition of resectability using three different factors: 1. anatomical; 2. biological; 3. conditional [7]. For anatomical criteria it basically includes the above mentioned factors and basically a serum carbohydrate antigen (CA) 19-9 level more than 500 U/mL or regional lymph node metastases diagnosed by biopsy or positron emission tomography-computed tomography. Potentially resectable disease based on anatomic criteria is transferred to borderline resectability if these factors are present. Conditional factors include the ECOG classification of patients and may also shift potentially resectable disease based on anatomic and biologic criteria towards a borderline resectable status if classification equals or exceeds ECOG 2 [7]. The detailed definition of anatomical resectability is displayed in Table 1. Recently, the BACAP Consortium published a BACAP Score to predict the resectability of pancreatic adenocarcinoma based on anatomical considerations (vascular thrombosis, tumor localization, tumor size) as well as conditional evaluation (WHO performance status) and symptoms (pain, weight loss) [16]. Furthermore, if the superior mesenteric artery or the celiac trunk are invaded, pancreatic tumors are considered as locally advanced and unresectable, however arterial resections and reconstructions can be performed by experienced surgeons. Actually, for determination of resectability, the relationship between the tumor and mesenteric/hepatic vessels is the critical topic to obtain R0 resection [13–15]. Hence, through preoperative staging and imaging, pancreatic tumors are divided into four types: resectable, borderline resectable, locally advanced and metastatic. Currently, upfront surgery is recommended in resectable pancreatic cancer [7,11,12]. In contrary to the surgical and anatomical considerations to evaluate resectability, the International Association of Pancreatology developed a more comprehensive definition of resectability using three different factors: 1. anatomical; 2. biological; 3. conditional [7]. For anatomical criteria it basically includes the above mentioned factors and basically a serum carbohydrate antigen (CA) 19-9 level more than 500 U/mL or regional lymph node metastases diagnosed by biopsy or positron emission tomography-computed tomography. Potentially resectable disease based on anatomic criteria is transferred to borderline resectability if these factors are present. Conditional factors include the ECOG classification of patients and may also shift potentially resectable disease based on anatomic and biologic criteria towards a borderline resectable status if classification equals or exceeds ECOG 2 [7]. The detailed definition of anatomical resectability is displayed in Table 1. Recently, the BACAP Consortium published a BACAP Score to predict the resectability of pancreatic adenocarcinoma based on anatomical considerations (vascular thrombosis, tumor localization, tumor size) as well as conditional evaluation (WHO performance status) and symptoms (pain, weight loss) [16]. Based on the analysis of a prospectively collected 814-patient cohort, this score will be evaluated in further clinical trials.

**Table 1.** International consensus of classification of BR PDAC based on anatomical definition using CT imaging including coronal and sagittal sections [7].


#: In cases with CA invasion of 180 or more without involvement of the aorta and with intact and uninvolved gastroduodenal artery thereby permitting a distal pancreatectomy with en bloc celiac axis resection (DP-CAR), some members prefer this criteria to be in the BR-A category. \$: including macroscopic para aortic and extra abdominal lymph node metastasis.

> Nowadays, CT, MRI, and PET are applied in the imaging detection and staging for patients with pancreatic cancer as well as endoscopic ultrasound (EUS). Contrast-enhanced CT is regarded as primary approach for the diagnosis and resectability evaluation. MRI is an alternative choice and is superior to CT when evaluating ductal anatomy with MRCP [17]. Besides, MRI is superior to detect the liver metastases compared to CT with higher sensitivity [18]. Recently, PET-MRI has been reported to have equal efficacy in resectability evaluation for the patients with PDAC [19]. Yamada et al. showed that EUS combined with elastography (EG) had better diagnostic performance in evaluating vascular invasion for PDAC compared to CT [20]. In addition, Ehrlich et al. also demonstrated that for patients with borderline resectable pancreatic cancer and locally advanced pancreatic cancer, EUS-FNA (fine-needle aspiration) has the potential to ensure the diagnose as well as local resectability accurately and suggested it as a routine approach for PDAC patients [21]. However, a recent meta-analysis indicated that CT might be superior to EUS in resectability evaluation; so a controversy about EUS application still remains and more high-quality clinical trials need to be conducted in the future to achieve more high-level evidence [22–24].

#### *1.2. Neoadjuvant and Adjuvant Therapy*

The impact of adjuvant chemotherapy to improve survival after resection of pancreatic cancer has been undoubtedly be proven during the last two decades, namely by the ESPAC study group as well as the PRODIGE consortium who continuously developed standards for adjuvant treatment by conducting large multicenter RCTs [25–28].

The latest of these studies reported median survival times of 30 and 54 months, respectively, as well as a 5-year survival of 30% which shows the essential need for adjuvant systemic treatment after pancreatic cancer resection [26,28]. This has ultimately been adopted in national and international guidelines [29].

Today there is a worldwide trend to increase the proportion of patients receiving neoadjuvant therapy. While neoadjuvant therapy is inevitable in locally advanced pancreatic cancers to achieve a chance of conversion surgery afterwards, its use in borderlineresectable and especially resectable pancreatic cancer is currently still based on weak

evidence, although observational and a limited number of randomized controlled trials suggest its benefit when borderline resectable disease is considered.

Yet, the main dilemma remains the selection of patients for neoadjuvant treatment and the selection of the specific treatment protocol. Neoadjuvant chemotherapy alone or in combination with radiotherapy is widely used in numerous varying protocols on one hand, on the other hand these protocols are often based on institutional or national preferences and—in contrast to adjuvant protocols—no standards are set on the basis of high-quality evidence [30,31].

Briefly, the debate on upfront surgery versus neoadjuvant treatment still remains. The Dutch Randomized Phase III PREOPANC Trial demonstrated that there was no significant difference in overall survival benefit between the preoperative chemoradiotherapy and upfront surgery for resectable and borderline resectable pancreatic cancer [32]. Given the observation that upfront surgery combined with adjuvant therapy can attain an average 19% five-year overall survival which increases up to 50% in prognostically favorable subgroups neoadjuvant therapy is still far from being the standard based on high-level evidence [33]. If neoadjuvant therapy is chosen, another unsolved question is the need for additional adjuvant therapy after resection. A recently published study pooling observational data of 520 patient after induction FOLFIRINOX treatment and consecutive resection showed that an additional adjuvant protocol did not generally show any benefit but may be recommended for pathologically lymph-node positive patients [34]. A phase 2 Randomized Clinical Trial discovered that perioperative chemotherapy did not significantly improve two-year overall survival for resectable PDAC whereas may increase actual resectability rates—an observation which is certainly explained by a selection effect during neoadjuvant treatment [35]. All in all, adjuvant and especially neoadjuvant treatment are currently in a dynamic state and numerous studies are ongoing.

#### **2. Surgery**

#### *2.1. Standard Resection*

Pancreaticoduodenectomy (PD) has been widely applied since in 1940, Whipple reported the classical procedure including distal gastrectomy and total duodenectomy and although this approach has been modified in some steps it is still basically similar to what is performed today [36].

PD includes a standardized lymphadenectomy along the right side of the vascular structures (porto-mesenteric veins, superior mesenteric artery, celiac axis) and the hepatoduodenal ligament. Nowadays, PD is routinely performed under preservation of the pylorus as recent studies have confirmed that pylorus preservation does not have any disadvantages compared to pylorus resection or classical Whipple procedures in terms of functional (especially regarding delayed gastric emptying) and oncological outcomes unless the tumor extends towards the pylorus, which then—unquestionably—requires resection of the distal stomach.

For tumors of the body and tail of the pancreas, a distal pancreatectomy and splenectomy with respective lymphadenectomy from the left side of the vascular structures is mandatory.

In case of unfavorable location of the tumor in the center of the pancreas or synchronous multiple PDAC, a total pancreatectomy and splenectomy may be required. Regarding all resection techniques, it is of the utmost importance to achieve a radical (R0) resection status. This can best be achieved by a complete dissection of all lymphatic and soft tissue along the arterial structures to reduce the risk of remaining microscopic tumor persistence and early recurrence.

#### *2.2. Specific Techniques*

#### 2.2.1. Artery First Approach

The core principle of this procedure is to identify the SMA early at the origin of the aorta and the approach has been described for different ways of access to the artery [37,38]. The idea of the approach to evaluate any potential tumor adherence to the SMA at the beginning of the operation and either stop resection or plan an arterial resection if required and indicated. After exposing the SMA from the left-sided access (opening Treitz ligament) a Kocher maneuver is required to expose the anterior surface of the inferior vena cava and the aorta with an early identification of the left renal vein and the origin of the SMA. After the accurate dissection along the SMA is finished, the soft tissue between the SMA and the celiac trunk should also be removed. By this procedure, a very controlled and radical resection on the right side of the arterial axis (SMA/celiac trunk) is achieved, while the autonomous nerves on the left side of the arteries are spared to reduce the incidence of postoperative diarrhea. For the radical resection of pancreatic head tumors which involve the posterior and right side of the SMA, the artery first technique is beneficial and recommended. A recent meta-analysis indicated that the SMA artery first approach can decrease the overall complication rate (OR 0.62, 95% 17 CI 0.47 to 0.81, *p* = 0.001) and reduce blood loss (WMD −264.84, 95% CI −336.1 to 18 −193.58, *p* < 0.001) compared to the normal procedure in pancreaticoduodenectomy and attain an increased R0 resection rate (OR 2.92, 95% CI 1.72 to 4.96, *p* < 0.001) and three-year OS (OR 2.15, 95% CI 1.34 to 3.43, *p* = 0.001) showing that the artery first approach can have superior clinical outcomes [39]. Until now, many different artery first approaches have been developed, such as the posterior approach, the right/medial uncinate approach, the inferior infracolic or mesenteric approach or the hanging maneuver [40]. This underlines the importance of paying attention to the status of the SMA and achieving an increased R0 rate through the meticulous dissection of the right margin of the SMA.

#### 2.2.2. Uncinate Process First

The Uncinate first approach describes a modified technique of resection along of the right margin of the SMV and SMA through a special method. This approach includes the division of the proximal jejunum and translocation of the first jejunal loop before other steps of dissection. Afterwards, the pancreatic head is dissected retrogradely and finally leading to the transection of the pancreas at its neck [41]. The first step of the approach is to open Treitz ligament from the left side of the mesenteric root after the Kocher maneuver with wide mobilization of the duodenum. After division and skeletonizing the first jejunal loop, this is then pulled through to the right side of the mesenteric root and resection can be continued as described above. When using this method, there is no need to use tunneling to transect the pancreas above the portal vein for the specimen is usually already mobilized extensively. Through the retrograde approach, the resection may be more radical due to a clear visualization of the medial resection margin throughout the entire preparation and both superior mesenteric vessels, arteries and veins are clearly seen which may reduce blood loss. Hence, it is recommended as an additional technique in modern pancreatic surgery. Recently, it was demonstrated that also laparoscopic uncinate first approach is a feasible method for pancreatic head neoplasms with high lymph node harvests (19.3 vs. 13.9 (*p* = 0.03)) and no significant difference in R0 resection, operative time and median length of stay compared to laparoscopic classical approach [42]. Zhang et al. reported that laparoscopic pancreaticoduodenectomy (LPD) combined with the uncinate process first approach improved the laparoscopic resection technique with low risk of postoperative complications and high rate of curative resection [43]. Wang et al. described that LPD with uncinate process first reduced the operative time, decreased the bleeding amount during the operation and protected the variant hepatic artery suggesting that it is safe and feasible to conduct LPD together with uncinate first approach [44]. Additionally, a recent comparative study displayed that LPD with the uncinate process-first approach was feasible compared to traditional pancreatic surgery for this new technique can achieve less blood loss and a shorter first flatus time together with diet start time [45].

#### 2.2.3. The TRIANGLE Operation cally advanced pancreatic cancer after the neoadjuvant therapy and was described in 2017

2.2.3. The TRIANGLE Operation

*Cancers* **2021**, *13*, x FOR PEER REVIEW 6 of 16

The TRIANGLE operation aims to develop a novel method for the patients with locally advanced pancreatic cancer after the neoadjuvant therapy and was described in 2017 [46]. The rationale of this procedure is the observation that after neoadjuvant therapy conventional imaging fails to differentiate between actual tumor encasement or abutment and only fibrotic residual tissue mainly to the arterial structures. Therefore, the technique comprises dissection of all soft tissue along the CA, SMA, SMV, and PV in association with a radical tumor removal. During the resection process, if must be proven that the specific periarterial tissue does not include viable tumor by frozen section; afterwards a radical artery-sparing approach can be conducted. This results in an anatomic triangle bordered by the SMA, CA, and portal vein revealed by the dissection and finally resection indicating the comprehensive removal of all soft tissue contained within these borders—usually fibrotic, neural, and lymphatic tissue (Figure 2). It is essential for the artery to be reached on the adventitial layer which opens longitudinally and allows to carry out the lymphadenectomy and soft tissue removal of the respective area. Above all, this technique allows patients after neoadjuvant therapy have the chance to attain a comprehensive tumor removal. Furthermore, the major advantage is the avoidance of arterial resection and reconstruction. [46]. The rationale of this procedure is the observation that after neoadjuvant therapy conventional imaging fails to differentiate between actual tumor encasement or abutment and only fibrotic residual tissue mainly to the arterial structures. Therefore, the technique comprises dissection of all soft tissue along the CA, SMA, SMV, and PV in association with a radical tumor removal. During the resection process, if must be proven that the specific periarterial tissue does not include viable tumor by frozen section; afterwards a radical artery-sparing approach can be conducted. This results in an anatomic triangle bordered by the SMA, CA, and portal vein revealed by the dissection and finally resection indicating the comprehensive removal of all soft tissue contained within these borders—usually fibrotic, neural, and lymphatic tissue (Figure 2). It is essential for the artery to be reached on the adventitial layer which opens longitudinally and allows to carry out the lymphadenectomy and soft tissue removal of the respective area. Above all, this technique allows patients after neoadjuvant therapy have the chance to attain a comprehensive tumor removal. Furthermore, the major advantage is the avoidance of arterial resection and reconstruction.

was feasible compared to traditional pancreatic surgery for this new technique can achieve less blood loss and a shorter first flatus time together with diet start time [45].

The TRIANGLE operation aims to develop a novel method for the patients with lo-

**Figure 2.** Intraoperative view after radical resection in pancreatic cancer (TRIANGLE operation). Porto-mesenteric vein resection and reconstruction with ringed allograft, dissection of all soft tissue (grey triangle) between celiac axis and superior mesenteric artery (red tapes) as well as the replaced mesenterico-portal vein. Blue tape: left kidney vein. **Figure 2.** Intraoperative view after radical resection in pancreatic cancer (TRIANGLE operation). Porto-mesenteric vein resection and reconstruction with ringed allograft, dissection of all soft tissue (grey triangle) between celiac axis and superior mesenteric artery (red tapes) as well as the replaced mesenterico-portal vein. Blue tape: left kidney vein.

Furthermore, the major advantage is the avoidance of arterial resection and reconstruction. However, when required, the TRIANGLE operation can be combined with and arterial resection and reconstruction, a venous resection is frequently required in this situation. Rosso et al. described that the "triangle operation" for borderline resectable pancreatic head cancer was safe and efficient [47].

#### 2.2.4. Venous Bypass First

One of the most challenging procedures during pancreatectomy can arise when venous infiltration of the portal/superior mesenteric vein axis is basically possible but hampered by large collateral vessels which implies that preparation may take a rather long time with the consecutive need for a long clamping time towards the small bowel with venous congestion [48]. In such situations, including cavernous transformation of the portal vein, a new surgical technique called "venous bypass graft first" is the procedure of choice [49,50]. The idea of this procedure is to create an initial venous bypass graft placement between the superior mesenteric vein or its tributaries and the portal vein in order to avoid bleeding as well as venous congestion of the small bowel. If the portal vein is not accessible in the hepatoduodenal ligament or liver hilum, this bypass can be performed between superior mesenteric vein and inferior cava vein after the Kocher/Cattel-Braasch maneuver is completed before proceeding with the resection of the pancreatic head. As cavernous transformation of the portal vein is caused by a complete portal/superior mesenteric vein occlusion; otherwise, it is an unsolved obstacle for resection, the stepby-step pancreatic head resections with a 'venous bypass graft first' approach should be carried out to overcome this problem. The approach includes preoperative assessment of the superior mesenteric and portal vein, exploration, and identification of venous vessels suitable for a graft placement. By this technique, a continuous porto-venous inflow to the liver during the resection phase is ensured if performed as a mesenterico-portal bypass. If this is not directly possible, at least a severe venous congestion of the small bowel can be avoided. in cases of temporary mesenterico-caval shunting and final restoration of the portal vein inflow reconstruction to the portal vein after completed tumor resection.

#### 2.2.5. Periarterial Divestment

Due to the increasing application of neoadjuvant therapy in PDAC, especially in locally advanced disease, surgical strategies and concepts have gradually changed as well as resection techniques, especially for cases which have been down-staged or shown a stable disease. It still remains controversial whether it is mandatory to perform arterial resection for arterial involvement in pancreatic cancer. An alternative approach has been described as the "periarterial divestment" technique [51,52]. This technique comprises a radical tumor clearance without arterial resection instead. Because of the inaccuracy of detection of true arterial involvement and true arterial invasion through current imaging methods, operative exploration should be performed.

The technique of periarterial divestment describes the sub-adventitial dissection in the layer between the arterial wall and remnant tumor/fibrous tissue which allows a radical removal without an arterial replacement. All in all, 'artery first' approach, 'uncinate process first', 'triangle operation', 'venous bypass first', and 'periarterial divestment' are complementary techniques in pancreatic cancer surgery. These mainly vessel-oriented technical approaches of pancreatic head resection allow removal of all putatively tumorinfiltrated soft tissue with the utmost aim for an improved R0 resection rate [53].

#### *2.3. Vascular (Venous and Arterial) Resection*

#### 2.3.1. Venous Resection

Vascular resection, especially for venous resection has now been widely applied with pancreaticoduodenectomy in selected patients. The earliest surgery focusing on the superior mesenteric vein (SMV) was reported by Moore in 1951 during pancreatic surgery [54]. Afterwards, en bloc pancreatoduodenectomy with vein resection was described by Fortner

and indicated that the technique is safe and favorable [55]. Venous resections have been modified and refined to be a routine surgical procedure in high volume centers [56,57]. It is possible to perform vein resection in patients with PDAC during all types of pancreatic surgery including pancreaticoduodenectomy, distal, or total pancreatectomies. The ISGPS classified mesentericoportal vein resections into four groups which was mainly considered by the approaches of resection and reconstruction [12]. Regarding outcomes of these techniques, vascular resection along with multiple treatments is beneficial for the patients with pancreatic cancer especially in the long-term overall survival. [58]. Several observational studies [59,60] demonstrated that neoadjuvant systematic chemotherapy can lead to increased radical resection chances for patients with complex tumor-vessel anatomy. The 2019 French Recommendations for the Vascular Resection for Pancreatic Cancer [60] has suggested that neoadjuvant treatment should be applied in case of venous tumor involvement followed by pancreatectomy with venous resection and can potentially be curative for the respective patients. It is unquestionable that venous resection during PD must also aim to obtain negative resection margins, while the reported effects on survival remain controversial [61]. A meta-analysis showed that pancreatectomy combined with venous resection needed longer operative time and had increased perioperative blood loss compared to the group of pancreatectomy without venous resection [62]. Patients with venous resection attained reduced R0 rates. There was no significant difference in postoperative complications between the two groups. In terms of survival, patients with venous resection had lower one-, three-, and five-year survival. The most recent meta-analysis [63] described that patients with pancreaticoduodenectomy plus venous resection seemed to attain a larger tumor size, positive lymph nodes and R1 resection rates and higher 30 day mortality. However, there was no significant difference in rates of total complications. In terms of long-term outcomes, patients with venous resection had lower one-year overall survival (OS), three-year OS, and five-year OS. A retrospective study [64] revealed that patients during pancreatic resection with venous vascular resection attained higher morbidity, lower five-year disease-free survival (7% and 20%, *p* = 0.018) and five-year disease-specific survival (19% and 35%, *p* = 0.42). Controversially to the reported impaired survival after venous resection, a recent propensity score-matched analysis [65] showed similar survival in pancreaticoduodenectomy with venous resection and pancreaticoduodenectomy alone groups after adjustment for baseline characteristics. A Japanese study [66] described the feasibility of venous resection and—in combination with adjuvant therapy—favorable outcomes reaching a 30-month median survival time in borderline resectable patients. This underlines the need to perform a venous resection whenever required to achieve negative resection margins and not to compromise radicality by avoidance of vascular resection and reconstruction. However, the effects of the various treatment options—including neoadjuvant therapy—in this setting require further evaluation and more high-level studies need to be conducted in the future.

#### 2.3.2. Artery Resection

In the 1950s, arterial resection was initially described during abdominal surgery by Appleby on resection of the celiac axis during extended gastrectomy including distal pancreatectomy [67]. In contrast to vein resection, artery resection is more debatable for its increased morbidity and mortality and mostly considered as an individual decision in selected patients [68]. However, the modified Appleby procedure which implies distal pancreatectomy, splenectomy, and celiac axis resection under preservation of the stomach has been shown to be beneficial for the patients with advanced tumors of the pancreatic body and tail [69]. This procedure can achieve median survival times of at least 18 months when combined with a multimodal treatment concept and is gaining increasing acceptance today [70]. Furthermore, during recent years, the techniques of replacement applied for the hepatic artery or the superior mesenteric artery have been improved and procedures such as splenic artery use have been described for restoration of hepatic or small-intestine perfusion (Figures 3 and 4) [71]. Oba et al. confirmed that arterial resection it is more likely

to attain preferable long-term outcome after the application of preoperative neoadjuvant treatments [72]. A Japanese study reported that patients with distal pancreatectomy plus celiac axis resection who underwent preoperative therapy achieved better one-, two-, and five-year overall survivals (100%, 90%, and 78.8%) than those who underwent upfront surgery (77.9%, 51.5%, and 26.7%; *p* < 0.0001) [73]. A recent meta-analysis showed that patients undergoing pancreatic surgery with artery resection had a greater risk of postoperative mortality (RR: 4.09, *p* < 0.001), morbidity (RR: 1.4, *p* = 0.01) and worse three-year survival [74]. Regarding specific complications and outcomes, the postoperative complications and the length of hospital stay and non-R0 rate were not significantly different compared to those without artery resection. A single-center cohort study reported that pancreatectomy with artery resection can attain better one-, three-, and five-year survival rates compared to palliation for patients with LAPC [75]. Another recent study covering nearly 40 years of experience showed that any type of arterial resection was performed at a frequency of 6% (44/730 patients) and confirmed the safety and efficacy of these operations for patients with locally advanced pancreatic cancer, additionally suggesting preoperative therapy with artery resection as a useful concept for locally advanced pancreatic cancer [76]. An important aspect in selecting patients properly and gaining sufficient surgical experience to safely perform such procedure which has recently been shown in two large series that demonstrated the impact of the surgical learning curve in two single center collectives of 111 and 195 patients, respectively [77,78]. createctomy plus celiac axis resection who underwent preoperative therapy achieved better one-, two-, and five-year overall survivals (100%, 90%, and 78.8%) than those who underwent upfront surgery (77.9%, 51.5%, and 26.7%; *p* < 0.0001) [73]. A recent meta-analysis showed that patients undergoing pancreatic surgery with artery resection had a greater risk of postoperative mortality (RR: 4.09, *p* < 0.001), morbidity (RR: 1.4, *p* = 0.01) and worse three-year survival [74]. Regarding specific complications and outcomes, the postoperative complications and the length of hospital stay and non-R0 rate were not significantly different compared to those without artery resection. A single-center cohort study reported that pancreatectomy with artery resection can attain better one-, three-, and fiveyear survival rates compared to palliation for patients with LAPC [75]. Another recent study covering nearly 40 years of experience showed that any type of arterial resection was performed at a frequency of 6% (44/730 patients) and confirmed the safety and efficacy of these operations for patients with locally advanced pancreatic cancer, additionally suggesting preoperative therapy with artery resection as a useful concept for locally advanced pancreatic cancer [76]. An important aspect in selecting patients properly and gaining sufficient surgical experience to safely perform such procedure which has recently been shown in two large series that demonstrated the impact of the surgical learning curve in two single center collectives of 111 and 195 patients, respectively [77,78].

small-intestine perfusion (Figures 3 and 4) [71]. Oba et al. confirmed that arterial resection it is more likely to attain preferable long-term outcome after the application of preoperative neoadjuvant treatments [72]. A Japanese study reported that patients with distal pan-

*Cancers* **2021**, *13*, x FOR PEER REVIEW 9 of 16

**Figure 3.** Example of splenic artery transposition on an aberrant right hepatic artery after resection of the aberrant hepatic artery due to tumor infiltration. Proper left hepatic artery with red tape and stump of the gastroduodenal artery (broken white arrow); portal vein (white asterisk); transposed splenic artery with end-to-end anastomosis on the aberrant right hepatic artery. **Figure 3.** Example of splenic artery transposition on an aberrant right hepatic artery after resection of the aberrant hepatic artery due to tumor infiltration. Proper left hepatic artery with red tape and stump of the gastroduodenal artery (broken white arrow); portal vein (white asterisk); transposed splenic artery with end-to-end anastomosis on the aberrant right hepatic artery.

**Figure 4.** Intraoperative view after combined arterial and venous resection during partial pancreato-duodenectomy. Resection of the common hepatic artery (white circle) and reconstruction by splenic artery transposition with end-to-end anastomosis (white arrow) on the proper hepatic artery (upper left red tape). Distal splenic artery stump (black circle) below the pancreatic cut margin; black asterisk: celiac axis; end-to-end reconstruction of the superior mesenteric/portal vein (dotted white arrow) and splenic vein on inferior mesenteric vein (black circle); upper right red tape: left gastric artery; lower middle red tape: superior mesenteric artery. **Figure 4.** Intraoperative view after combined arterial and venous resection during partial pancreato-duodenectomy. Resection of the common hepatic artery (white circle) and reconstruction by splenic artery transposition with end-to-end anastomosis (white arrow) on the proper hepatic artery (upper left red tape). Distal splenic artery stump (black circle) below the pancreatic cut margin; black asterisk: celiac axis; end-to-end reconstruction of the superior mesenteric/portal vein (dotted white arrow) and splenic vein on inferior mesenteric vein (black circle); upper right red tape: left gastric artery; lower middle red tape: superior mesenteric artery.

#### *2.4. Multivisceral Resection 2.4. Multivisceral Resection*

In addition to vascular resections, also multivisceral resections have been applied increasingly nowadays to attempt to achieve margin-negative resection. Previous studies indicate that pancreaticoduodenectomy with multivisceral resection is associated with increased morbidity and potentially mortality with conflicting results in terms of oncologic outcomes [79–81]. A systematic review suggested that multivisceral pancreatectomies was safe and feasible in selected patients [82]. A case-matched study showed that multivisceral distal pancreatectomy was able to achieve radical tumor removal providing beneficial survival outcomes [83]. Furthermore, a single center analysis demonstrated that multivisceral resection in pancreatic surgery was suitable for locally advanced pancreatic carcinoma of the body and/or tail [84], comparable results were achieved in a current multi-center publication, proving that distal pancreatectomy with multivisceral resection is viable in order to obtain free margins which is the key to achieve long-term survival [85]. In addition to vascular resections, also multivisceral resections have been applied increasingly nowadays to attempt to achieve margin-negative resection. Previous studies indicate that pancreaticoduodenectomy with multivisceral resection is associated with increased morbidity and potentially mortality with conflicting results in terms of oncologic outcomes [79–81]. A systematic review suggested that multivisceral pancreatectomies was safe and feasible in selected patients [82]. A case-matched study showed that multivisceral distal pancreatectomy was able to achieve radical tumor removal providing beneficial survival outcomes [83]. Furthermore, a single center analysis demonstrated that multivisceral resection in pancreatic surgery was suitable for locally advanced pancreatic carcinoma of the body and/or tail [84], comparable results were achieved in a current multi-center publication, proving that distal pancreatectomy with multivisceral resection is viable in order to obtain free margins which is the key to achieve long-term survival [85].

#### *2.5. MIS/Robotic Surgery 2.5. MIS/Robotic Surgery*

With the rapid development of technology, minimal invasive pancreatic surgery has been popularly applied worldwide. The first laparoscopic pancreatectomy was reported in 1996 while the first robotic pancreatic resections were described in 2003 [86,87]. A recent international evidence-based guideline on minimally-invasive pancreatic surgery demonstrated that open, laparascopic and robotic pancreatic surgery all have their own aspects in treating patients with pancreatic diseases and it is quite possible to achieve promising clinical outcomes by applying these advanced technologies [88]. An international consensus statement on robotic pancreatic surgery showed that robotic pancreatic surgery is safe and feasible compared to open pancreatic surgery [89]. Another international expert consensus on laparoscopic pancreaticoduodenectomy (PD) also showed that laparoscopic With the rapid development of technology, minimal invasive pancreatic surgery has been popularly applied worldwide. The first laparoscopic pancreatectomy was reported in 1996 while the first robotic pancreatic resections were described in 2003 [86,87]. A recent international evidence-based guideline on minimally-invasive pancreatic surgery demonstrated that open, laparascopic and robotic pancreatic surgery all have their own aspects in treating patients with pancreatic diseases and it is quite possible to achieve promising clinical outcomes by applying these advanced technologies [88]. An international consensus statement on robotic pancreatic surgery showed that robotic pancreatic surgery is safe and feasible compared to open pancreatic surgery [89]. Another international expert consensus on laparoscopic pancreaticoduodenectomy (PD) also showed that laparoscopic

pancreaticoduodenectomy was safe and effective for experienced surgeons [90]. Furthermore, a current network meta-analysis indicated that laparoscopic PD and robotic PD had a reduced length of hospital stay, operative bleeding and overall complications while on the other hand achieving a similar number of retrieved lymph nodes, tumor-free resection margins, clinically relevant postoperative pancreatic fistula, severe postoperative complications [91]. Besselink et al. demonstrated that minimally invasive distal pancreatectomy (DP) is technically safe, whereas oncological feasibility needs to be evaluated carefully. With respect to minimally invasive PD, some advantages have been shown in comparison to open PD [92]. However, due to a limited level of evidence, this has to be regarded with care, but minimally-invasive PD could be beneficial for selected patients with better short-term clinical outcomes. Without doubt, there is a strong need for more high-quality trials to confirm potential advantages of minimally invasive pancreatic surgery.

#### 2.5.1. Laparoscopic and Robotic Distal Pancreatectomy

The latest study indicated a shorter length of hospital, less delayed gastric emptying, higher rates of postoperative pancreatic fistula in minimally invasive distal pancreatectomy in contrast to open distal pancreatectomy [93]. The DIPLOMA study indicated that minimally invasive distal pancreatectomy attained less median blood loss, shorter hospital stay and less lymph node retrieval [94]. Furthermore, the LEOPARD randomized controlled trial proved that the less operative blood loss and the rate of delayed gastric emptying. However, longer operation times (217 vs. 179 min, *p* = 0.005) were observed in minimally invasive distal pancreatectomy [95]. A multicenter study described no significant differences in the incidence of clinically relevant postoperative pancreatic fistula in robotic distal pancreatectomy in contrast with open pancreatectomy [96].

#### 2.5.2. Laparoscopic and Robotic Pancreatoduodenectomy

A recent study showed that laparoscopic pancreaticoduodenectomy had lower blood loss, longer operative time. However, there were no obvious differences among 90-day overall mortality, Clavien-Dindo 3 complications and postoperative length of hospital stay in contrast to open surgery [97]. This meta-analysis included—among two other studies the LEOPARD-2 randomized controlled phase 2/3 trial which reported more complicationrelated deaths in the laparoscopic group compared to open pancreaticoduodenectomy with no obvious difference in time to functional recovery between the two groups and thereby weakened the conclusion that laparoscopic pancreaticoduodenectomy is potentially harmful [98]. Yet, this procedure is probably only feasible in highly-specialized centers with a respective high case load. However, the level of the current evidence focusing on minimal invasive pancreaticoduodenectomy may be too low, hence, more high-quality studies need to be carried out to enhance the future evidence, as especially for robotic procedures no randomized controlled trials are available to date.

#### **3. Conclusions and Future Perspective**

Over the past decades, surgical therapy for pancreatic cancer has been changing and developing rapidly allowing extended resections and improved complication. Given the advanced technology and comprehensive strategies, approaches of curative resections have improved as well as the quality of perioperative management. As a result, the mortality rate after pancreatic surgery has reduced obviously to a current rate of less than 5% in specialized centers. Although centralization has not become reality in all countries around the world, this should be the benchmark and especially advanced pancreatic surgery should be clearly limited to high-volume centers. Combined with multimodal treatment, pancreatic surgery allows to improve the quality of life and long-term survival for patients in different stages of pancreatic cancer. In terms of surgical techniques, open, laparoscopic and robotic procedures all will exert their own merit in their particular field to achieve benefit for the patients at the greatest extent.

**Author Contributions:** K.W.: study concept and design, drafting of the manuscript. T.H.: idea, study supervision, tables and figures preparation, critical revision of the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable as no humans or animals were involved in this study.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


## *Review* **Radical Resection for Locally Advanced Pancreatic Cancers in the Era of New Neoadjuvant Therapy—Arterial Resection, Arterial Divestment and Total Pancreatectomy**

**Yosuke Inoue \* , Atushi Oba, Yoshihiro Ono, Takafumi Sato, Hiromichi Ito and Yu Takahashi**

Division of Hepatobiliary and Pancreatic Surgery, Cancer Institute Hospital, Japanese Foundation for Cancer Research, Tokyo 135-8550, Japan; atsushi.oba@jfcr.or.jp (A.O.); yoshihiro.ono@jfcr.or.jp (Y.O.); takafumi.sato@jfcr.or.jp (T.S.); hiromichi.ito@jfcr.or.jp (H.I.); yu.takahashi@jfcr.or.jp (Y.T.) **\*** Correspondence: yosuke.inoue@jfcr.or.jp; Tel.: +81-3-3520-0111; Fax: +81-3-3570-0343

**Simple Summary:** Aggressive arterial resection or total pancreatectomy in surgical treatment for locally advanced pancreatic cancer (LAPC) has gradually been encouraged thanks to new chemotherapy regimens such as FOLFIRINOX or Gemcitabine and nab-paclitaxel, which have provided more adequate patient selection and local tumor suppression, justifying aggressive local resection. The development of surgical techniques provides the safety of arterial resection (AR) for even major visceral arteries, such as the celiac axis or superior mesenteric artery. Total pancreatectomy has been reevaluated as an effective option to balance both the local control and postoperative safety. In this review, we investigate the recent reports focusing on arterial resection and total pancreatectomy for locally advanced pancreatic cancer (LAPC) and discuss the rationale of such an aggressive approach in the treatment of PC.

**Abstract:** Aggressive arterial resection (AR) or total pancreatectomy (TP) in surgical treatment for locally advanced pancreatic cancer (LAPC) had long been discouraged because of their high mortality rate and unsatisfactory long-term outcomes. Recently, new chemotherapy regimens such as FOLFIRINOX or Gemcitabine and nab-paclitaxel have provided more adequate patient selection and local tumor suppression, justifying aggressive local resection. In this review, we investigate the recent reports focusing on arterial resection and total pancreatectomy for LAPC and discuss the rationale of such an aggressive approach in the treatment of PC. AR for LAPCs is divided into three, according to the target vessel. The hepatic artery resection is the simplest one, and the reconstruction methods comprise end-to-end, graft or transposition, and no reconstruction. Celiac axis resection is mainly done with distal pancreatectomy, which allows collateral arterial supply to the liver via the pancreas head. Resection of the superior mesenteric artery is increasingly reported, though its rationale is still controversial. Total pancreatectomy has been re-evaluated as an effective option to balance both the local control and postoperative safety. In conclusion, more and more aggressive pancreatectomy has become justified by the principle of total neoadjuvant therapy. Further technical standardization and optimal neoadjuvant strategy are mandatory for the global dissemination of aggressive pancreatectomies.

**Keywords:** pancreatic cancer; arterial resection; total pancreatectomy; neoadjuvant therapy

#### **1. Introduction**

Pancreatic cancer (PC) is a dismal clinical entity [1]. For localized PCs, resection is the only chance for cure. Theoretically, R0 resection is one essential philosophy for cancer treatment even if the local tumor has invaded major visceral arteries. However, the aggressive biology of PC accompanied with occult metastasis has precluded simply extending the resection. Pancreatectomy is accompanied by high morbidity, and extended resection, including arterial resection (AR) or multi-organ resection, has been a challenge because

**Citation:** Inoue, Y.; Oba, A.; Ono, Y.; Sato, T.; Ito, H.; Takahashi, Y. Radical Resection for Locally Advanced Pancreatic Cancers in the Era of New Neoadjuvant Therapy—Arterial Resection, Arterial Divestment and Total Pancreatectomy. *Cancers* **2021**, *13*, 1818. https://doi.org/10.3390/ cancers13081818

Academic Editor: Sohei Satoi

Received: 10 February 2021 Accepted: 7 April 2021 Published: 10 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of its substantial mortality [2,3]. Total pancreatectomy (TP) is an option to achieve R0 resection in locally advanced PCs. The rationale of TP for PC, however, has long been in controversy due to complicated short-time outcomes, including malnutrition or brittle diabetes, along with unsatisfactory long-term survival [4].

New-generation chemotherapies, i.e., FOLFIRINOX [5] or gemcitabine (GEM) + nabpaclitaxel (GNP) [6], have changed the paradigm of the treatment strategy for unresectable locally advanced (LA) PCs. In this review, we investigate the recent innovation of aggressive resection for LAPCs including AR or TP and discuss the future perspective of extended resections for advanced PCs.

#### **2. Arterial Resections**

#### *2.1. Overview*

Pancreatic ductal adenocarcinoma has an invasive nature, and our predecessor surgeons have tried to improve the prognosis by achieving R0 resection by extending resection. Fortner et al. reported the first series of extended resections named regional pancreatectomy [7]. This report described a novel approach of pancreatectomy for PCs, including TP and routine portomesentericosplenic confluence resection en bloc with the surrounding soft tissue. AR was concomitantly performed if needed. However, their results showed severe short-term outcomes and insufficient long-term survivals and was not accepted as a reasonable method to improve the treatment outcomes of LAPCs [2,7]. Since then, advances in surgical techniques and perioperative management have made venous resection and reconstruction during pancreatectomy safe [8,9]. Recent reports have documented favorable short-term outcomes of venous resection in patients with localized PCs [10,11]; however, the R0 resection rate, as well as long-term survival, remained unsatisfactory, because the most frequent site of cancer-positive margin was located at the superior mesenteric artery (SMA) margin [12,13], which could not be overcome by venous resection alone. Therefore, the necessity of more radical dissection, including arterial resection, remained and has become more prominent in the past two decades, although recent meta-analyses concluded that pancreatectomy with ARs remained a challenge, because it increased the complexity of the procedure and was associated with increased morbidity and mortality in comparison to non-AR pancreatectomies [14,15].

#### *2.2. Management for the Involvement of the Superior Mesenteric Artery*

In advanced pancreatic uncinate cancers, the superior mesenteric artery (SMA) is the most common artery that is invaded and becomes a reason for unresectable or pathologically noncurative resection [12,13,16]. Until recent years, a large series of SMA resections for PCs was quite limited, and mortality after SMA resection had reportedly been higher than ordinal pancreatectomies, which discouraged the aggressive resection of LAPCs involving the SMA [17–20] (Table 1). As an alternative, periadventitial dissection (PAD) of the SMA had been proposed to pursue the local control of the peri-SMA region. Inoue et al. presented a standardized technique of SMA-PAD using the supracolic anterior artery-first approach, which resulted in no mortality over 158 patients, with a R0 rate of 74% [16,21]. Extended resection of the peri-SMA nerve plexus was assumed to cause neurogenic diarrhea, which would lead to insufficient patient recovery or adjuvant therapy. Inoue et al. documented that the incremental administration of an opium tincture according to the frequency of watery diarrhea was effective and easy to adjust to, with satisfactory diarrhea control, leading to sufficient adjuvant therapy introduction (83%) [16]. For more advanced tumors that cause encasement of the artery, SMA resection would be required. Recently, some high-volume centers with outstanding expertise in pancreatic resections have reported large series of arterial resections for PCs, including more than 30 cases of SMA resections [22,23]. Bachellier et al. [22] reported a large single-center series, including 34 SMA resections. They achieved the lowest mortality ever (5.1% of all patients with AR), which represented the improved safety of SMA resection and reconstruction. They mainly employed an end-to-end anastomosis using autografts such as a great saphe-

nous vein and noted that reconstruction with an artificial graft caused thrombosis, leading to in-hospital mortality. Loos et al. [23] reported another large series involving 30 SMA resections with an acceptable mortality of 6.7%. They also performed a learning curve analysis and concluded that even an experienced pancreatic surgeon needed more than 12 cases of AR to minimize the mortality. An optimal reconstruction technique has never been established and likely depends on the length of a resected segment. Previous reports on SMA reconstruction employed end-to-end anastomosis or anastomosis to the aorta with or without graft interposition (Figure 1A,B,D,E) and a rotation of the splenic artery (SpA) (Figure 1C) [10,17,20,22–34]. Westermark et al. [35] proposed a safe technique of end-to-end anastomosis of the SMA. They recommended the Cattel-Braasch maneuver, wherein the total mesentery is detached from the retroperitoneum to facilitate a tension-free anastomosis. Sterile ice in a surgical towel was placed in the lower sub-mesocolic abdomen to reduce the warm ischemia of the small intestine. The Cattel-Braasch maneuver enabled tension-free anastomosis even after SMA resection of 4 cm in length. Accordingly, SMA resection is now no more an anecdotal tool but one possible option for LA pancreatic head cancers. Reports focusing on the long-term outcomes after SMA resection are still limited. *Cancers* **2021**, *13*, x 4 of 15

**Figure 1.** Reconstruction of the superior mesenteric artery. (**A**) Basic anatomy of relevant vessels in SMA resection. (**B**) Direct end-to-end anastomosis. (**C**) Transposition of SpA to be anastomosed with the distal stump of the SMA. (**D**) Endto-end anastomosis with graft interposition. (**E**) Graft interposition from the aorta to the distal stump of the SMA. (**F**) Combined resection and reconstruction of the HA and SMA using interposition grafts. HA, hepatic artery, SpA, splenic artery, GDA, gastroduodenal artery, SMA, superior mesenteric artery, MCA, middle colic artery and LGA, left gastric artery. **Figure 1.** Reconstruction of the superior mesenteric artery. (**A**) Basic anatomy of relevant vessels in SMA resection. (**B**) Direct end-to-end anastomosis. (**C**) Transposition of SpA to be anastomosed with the distal stump of the SMA. (**D**) End-to-end anastomosis with graft interposition. (**E**) Graft interposition from the aorta to the distal stump of the SMA. (**F**) Combined resection and reconstruction of the HA and SMA using interposition grafts. HA, hepatic artery, SpA, splenic artery, GDA, gastroduodenal artery, SMA, superior mesenteric artery, MCA, middle colic artery and LGA, left gastric artery.

Advanced cancers located at the pancreatic neck often invade the common and proper hepatic artery (HA), as well as the gastroduodenal artery (GDA). In such cases, segmental resection of the HA, including the root of the GDA, is suggested. If cancer invasion is limited and resected segment is short, end-to-end anastomosis is often possible. Recent guidelines have also described the combined HA or celiac axis (CA) resection as one of the putative options for LAPC [36]. Although a large series that specifically focuses on HA resection is limited, there are many small case series, including five to 20 patients who mainly underwent pancreaticoduodenectomy (PD) with concomitant resection of the

outcomes about HA reconstruction. The in-hospital mortality rate was 7%, accompanied by an R0 rate of 80% and a median survival time (MST) of 12 months. The authors concluded that HA resection is justified only when surgery of R0 has taken place for selected patients with PC. Regarding the reconstruction technique of the HA, several reports described HA reconstruction, which was dominantly done by end-to-end anastomosis (Figure 2A,B) [24–27,30–32,37]. Short-segment resection of the HA was simple and safe and could be recommended as an entry procedure of AR for pancreatic surgeons who perform pancreatic head resection. In a case where the HA is resected in a long segment, arterial transposition (Figure 2C,D) [17,32] or interposition using the autograft to bridge between the celiac axis or aorta and proper HA is required (Figure 2E) [24,25,27,31]. To simplify and reduce the number of anastomoses, transposition of the SpA or colic artery should first be considered. The right inferior phrenic artery is an alternative option for a small orifice of the left HA. Although SpA transposition is usually performed with TP to gain enough length of the SpA pedicle, preservation of the pancreas tail would be possible if

*2.3. Resection of the Hepatic Artery* 


**Table 1.** Previous reports about resection of the superior mesenteric artery.

NAT, neoadjuvant therapy, ND, not described, PD, pancreaticoduodenectomy, TP, total pancreatectomy, DP, distal pancreatectomy, EEA, end-to-end anastomosis and SpA, splenic artery. \* The hepatic artery was anastomosed to the SpA with total pancreatectomy; \*\* Incidence among all patients with arterial resection. † Each number was not documented.

#### *2.3. Resection of the Hepatic Artery*

Advanced cancers located at the pancreatic neck often invade the common and proper hepatic artery (HA), as well as the gastroduodenal artery (GDA). In such cases, segmental resection of the HA, including the root of the GDA, is suggested. If cancer invasion is limited and resected segment is short, end-to-end anastomosis is often possible. Recent guidelines have also described the combined HA or celiac axis (CA) resection as one of the putative options for LAPC [36]. Although a large series that specifically focuses on HA resection is limited, there are many small case series, including five to 20 patients who mainly underwent pancreaticoduodenectomy (PD) with concomitant resection of the HA until recently [10,17,20,24–27,29–32,37–39] (Table 2). Amano H et al. first reported a medium series of HA resections in which they described the details of techniques and outcomes about HA reconstruction. The in-hospital mortality rate was 7%, accompanied by an R0 rate of 80% and a median survival time (MST) of 12 months. The authors concluded that HA resection is justified only when surgery of R0 has taken place for selected patients with PC. Regarding the reconstruction technique of the HA, several reports described HA reconstruction, which was dominantly done by end-to-end anastomosis (Figure 2A,B) [24–27,30–32,37]. Short-segment resection of the HA was simple and safe and could be recommended as an entry procedure of AR for pancreatic surgeons who perform pancreatic head resection. In a case where the HA is resected in a long segment, arterial transposition (Figure 2C,D) [17,32] or interposition using the autograft to bridge between the celiac axis or aorta and proper HA is required (Figure 2E) [24,25,27,31]. To simplify and reduce the number of anastomoses, transposition of the SpA or colic artery should first be considered. The right inferior phrenic artery is an alternative option for a small orifice of the left HA. Although SpA transposition is usually performed with TP to gain enough length of the SpA pedicle, preservation of the pancreas tail would be possible if the left gastric artery (LGA) and great pancreatic artery are preserved. Desaki et al. reported a case series of SpA resection during PD mainly for PCs and documented that no clinically

relevant splenic infarction was observed [40]. On the other hand, the omittance of HA reconstruction would be possible if we performed a specific preparation for HA resection. Miyazaki et al. proposed the novel management of HA resection with preoperative HA embolization to enhance the collateral hepatic arterial inflow [38]. After HA resection, backflow from the proper HA stump was observed. If the backflow was strong enough, they omitted HA reconstruction. In a 21-patient series, they reconstructed HA in only one patient, and eventually, 33% of the patients suffered postoperative liver infarction, but there was no in-hospital mortality.


**Table 2.** Previous reports about resection of the hepatic artery.

EEA, end-to-end anastomosis, ND, not described, PD, pancreaticoduodenectomy, TP, total pancreatectomy, DP, distal pancreatectomy, GDA, gastroduodenal artery and SpA, splenic artery. \* The replaced hepatic artery was anastomosed to the GDA. \*\* The hepatic artery was anastomosed to the SpA with total pancreatectomy. † Incidences among all patients with arterial resection. ‡ Includes patients who had replaced HA. § Each number was not documented.

with arterial resection. ‡ Includes patients who had replaced HA. § Each number was not documented.

tion, but there was no in-hospital mortality.

Amano H [17] 2009 Japan 15 13 † PD, TP EEA 3, GDA 4 \*

**Table 2.** Previous reports about resection of the hepatic artery.

Li [24] 2004 China 8 ND PD EEA 5, graft 3 1994–2003 ND Nakao [10] 2006 Japan 9 ND PD, TP ND 1981–2005 ND Yekebas [20] 2008 Germany 10 ND PD, TP, DP EEA 10 1994–2005 0

Boggi [25] 2009 Italy 12 ND PD EEA 6, graft 5, no reconstruction 1 1987–2004 4 † Martin [26] 2009 USA 3 33 PD, TP EEA 3 1999–2007 0 Bockhorn [27] 2011 Germany 18 ND PD, TP EEA 10, graft 8 1994–2004 14 † Gong [29] 2013 China 5 ND PD ND 2006–2011 6.7 † Amano R [37] 2015 Japan 7 100 PD, TP EEA 6, no reconstruction 1 2012–2013 0 Sgroi [30] 2015 USA 7 38 † PD EEA 7 2003–2013 ND Glebova [31] 2016 USA 18 28 † PD EEA 15, graft 2, no reconstruction 1 1989–2014 ND Perinel [32] 2016 France 6 0 TP SpA 3, no reconstruction 3 ‡ 2008–2014 0 Miyazaki [38] 2017 Japan 21 43 PD, TP EEA1, no reconstruction 20 2019–2015 0 Tee [33] 2018 USA 60 75 † PD, DP, TP EEA or graft or reconstruction § 1990–2017 13 Loveday [34] 2019 Canada 10 94 † PD, DP, TP EEA, interposition from the aorta † 2009–2016 3.2 † Bachellier [22] 2020 France 29 75 † PD, DP, TP EEA or graft 20 §, no reconstruction 9 § 1990–2017 5.1 † Loos [23] 2020 Germany 85 49 † PD, DP, TP EEA, graft, transposition § 2003–2019 16.7 EEA, end-to-end anastomosis, ND, not described, PD, pancreaticoduodenectomy, TP, total pancreatectomy, DP, distal

**Author Year Country** *N* **NAT (%) Procedures Reconstruction Method Study Period Mortality** 

the left gastric artery (LGA) and great pancreatic artery are preserved. Desaki et al. reported a case series of SpA resection during PD mainly for PCs and documented that no clinically relevant splenic infarction was observed [40]. On the other hand, the omittance of HA reconstruction would be possible if we performed a specific preparation for HA resection. Miyazaki et al. proposed the novel management of HA resection with preoperative HA embolization to enhance the collateral hepatic arterial inflow [38]. After HA resection, backflow from the proper HA stump was observed. If the backflow was strong enough, they omitted HA reconstruction. In a 21-patient series, they reconstructed HA in only one patient, and eventually, 33% of the patients suffered postoperative liver infarc-

**(%)** 

, SpA 6 \*\*, Others 3 2005–2009 6.7

**Figure 2.** Reconstruction of the hepatic artery. (**A**) Basic anatomy of the relevant vessels in HA resection. (**B**) Direct endto-end anastomosis. (**C**) Transposition of the MCA and RIPA to be anastomosed with the RHA and LHA. (**D**) Transposition of the SpA to be anastomosed with the proper HA. (**E**) Graft interposition from the aorta to the stump of the proper HA. HA, hepatic artery, RIPA, right inferior phrenic artery, SpA, splenic artery, GDA, gastroduodenal artery, SMA, **Figure 2.** Reconstruction of the hepatic artery. (**A**) Basic anatomy of the relevant vessels in HA resection. (**B**) Direct end-toend anastomosis. (**C**) Transposition of the MCA and RIPA to be anastomosed with the RHA and LHA. (**D**) Transposition of the SpA to be anastomosed with the proper HA. (**E**) Graft interposition from the aorta to the stump of the proper HA. HA, hepatic artery, RIPA, right inferior phrenic artery, SpA, splenic artery, GDA, gastroduodenal artery, SMA, superior mesenteric artery, MCA, middle colic artery, LGA left gastric artery, RHA, right hepatic artery and LHA, left hepatic artery.

#### *2.4. Resection of the Celiac Axis*

CA resection for advanced pancreatic body cancer was an exceptional situation of arterial resection, wherein reconstruction of the hepatic artery was considered to be unnecessary because of the peripancreatic collateral arterial flow that originated from the SMA [41]. Pancreatic body cancers frequently involve the celiac–hepatic artery system, and distal pancreatectomy with celiac axis resection (DP-CAR) was a reasonable choice to achieve an en-bloc eradication of the tumor and its invasion. The concept of DP-CAR was a modification of the Appleby procedure originally for advanced gastric cancers [41]. The first report about DP-CAR was written by Hishinuma et al. in 1991, and they documented the preservation of the whole stomach during CAR and distinguished DP-CAR from the Appleby procedure in that the stomach was preserved [42]. Afterward, several small series of DP-CARs were reported [43–47], and in 2007, Hirano et al. first described the short- and long-term outcomes of the standardized DP-CAR [48]. They reported 23 patients who underwent DP-CARs with no mortality and had acceptable overall survival (five-year survival rate, 42% and median survival time, 21 months). This pivotal report encouraged pancreatic surgeons worldwide to perform DP-CAR as a promising option to balance surgical and oncological safety. However, as the cases accumulated, ischemic complications involving the stomach or liver became prominent, as well as post-pancreatectomy hemorrhage, caused by the insufficient drainage of postoperative pancreatic fistula, leading to non-negligible mortality [49–54] (Table 3). Ischemic gastropathy or stomach perforation were complications specific to DP-CARs, which often included resection of the LGA, as well as the left gastroepiploic artery. Moreover, radical retroperitoneal dissection during DP-CAR includes resection of the left inferior phrenic artery. These sacrifices of critical gastric inflows potentially lead to life-threatening gastropathy [55]. As for liver infarction, collateral hepatic flow via the GDA was theoretically sufficient for liver perfusion. However, excessive dissection of the GDA sometimes leads to arterial stenosis, which causes depression of the hepatic arterial flow [56]. Depression of the proper hepatic artery induces recurrent cholangitis, liver abscess or cholecystitis. Cholecystitis was reported to be one possible cause of postoperative major intervention [50,55]. Therefore, the gallbladder should be resected routinely during DP-CAR. In the early years, preoperative arterial embolization of the HA or LGA to enhance the collateral flow was encouraged to avoid ischemic complications. However, recent reports found no positive impact of arterial embolization on the prevention of postoperative ischemic complications [55–58]. Another possible resolution is an intraoperative reconstruction of the LGA. Sato et al. first described reconstruction of the LGA to avoid ischemic gastropathy after DP-CAR [59]. The authors

used a pedicle of the middle colic artery as an origin of the arterial supply. The right branch of the middle colic artery is usually away from the pancreatic body cancer and used as a suitable counterpart of the LGA. The efficacy of the anastomosis should be confirmed promptly and objectively after anastomosis. Oba et al. reported the intraoperative evaluation of the patency of LGA anastomosis using indocyanine green fluorescence imaging [60]. By these managements, the safety of DP-CARs would be improved.

**Table 3.** Previous reports of distal pancreatectomy with celiac axis resections (DP-CARs).


LGA, left gastric artery and ND, not described.

#### **3. Total Pancreatectomy**

TP was reported by Rockey et al. for the first time [61]. Although TP was attempted to improve the survival of patients with PC with the rationale to avoid anastomosis-related morbidity and mortality in early years [62,63], Warren et al. documented that TP led to pancreatic endocrine and exocrine insufficiency, resulting in brittle diabetes due to a lack of endocrine and malabsorption caused by exocrine deficiency [64]. Later, TP was indicated with the intention to improve the local control in extensive pancreatic cancers [7]. However, as was described by Fortner et al., a simple extension of resection resulted in poor short-term outcomes accompanied by unsatisfactory survival [2]. In the 1980s, TP was attempted to eradicate multicentric carcinogenesis in the whole pancreas, but it failed to improve the survival of patients with PS, because the incidence of tumor multicentricity proved to be low [4,65]. Therefore, TP has been discouraged for the curative treatment of PCs [66]. After 2000, the introduction of long-acting insulin formulations facilitated the easy control of blood sugar levels after TP. As a result, endocrine-related mortality has been rarely reported ever since. As for exocrine insufficiency, diarrhea was the most frequent sequelae after TP, and 23.5% of patients who underwent TP still had symptoms, despite pancreatic enzyme administration [67]. Moreover, malabsorption causes postoperative steatohepatitis, which potentially leads to life-threatening hepatic decompensation [68]. Hata et al. identified female gender, malnutrition and insufficient pancreatic enzyme substitution as significant prognostic factors of post-TP steatohepatitis and suggested that high-dose pancreatic enzyme replacement therapy might have preventive effects on hepatic steatosis occurring after a pancreatectomy [69]. Anyway, the development and standardization of the surgical technique fostered by the case accumulation and centralization of complicated procedures has gradually made the surgical outcomes of TP an acceptable level, like partial pancreatectomy [70–73]. Long-term survivals have gradually become better and better. Until the middle of the 2000s, the MST of patients who underwent TP for PCs was about one year or less [4,74,75]. Schmidt et al. reported a substantial improvement in survival after TP for pancreatic neck cancers, documenting an MST of 18 months [76]. After 2010, a large series comprising 289 patients with TP for PCs documented an MST of 18.1 months [72]. Accordingly, TP was gradually reappraised as a reasonable option to achieve a cure for selective patients with PC [70,72,77–83] (Table 4). If TP was applied to LAPCs to obtain a cure or long-term survival, we would have to consider the quality of life after TP, as well as the absolute surgical safety or survival time.

Recently, several reports documented a significant reduction of physical functioning [84] or both the physical and emotional composite scores [85,86]. Stoop et al. stated in the latest report that the quality of life after TP was reduced in comparison to the general population but remained stable compared with the preoperative situation [84]. They demonstrated the challenges of endocrine (96% of patients involved) and exocrine insufficiency (64% of patients involved) after TP and claimed that the management of both insufficiencies should be improved further to overcome the quality of life reduction after TP.


**Table 4.** Previous reports about total pancreatectomy for pancreatic cancers.

ND, not described.

#### **4. Recent Evolution of Radical Pancreatectomies in the Era of New Regimens and Future Perspective**

#### *4.1. Recent Reports of Extremely Radical Pancreatectomy*

The respective techniques of arterial resection or total pancreatectomy have gradually matured and become common among experienced pancreatic surgeons; however, extremely radical pancreatectomy involving major arterial resection with or without total pancreatectomy is still controversial in that long-term survival is not considered worth carrying the surgical risks for patients with LAPC [3,17,70–72]. However, the introduction of new-generation chemotherapy regimens such as FOLFIRINOX [5] or GNP [6] has gradually changed the paradigm of indication for these surgical challenges. In recent years, multiple high-volume pancreatic centers have reported extremely radical pancreatectomy after intensive neoadjuvant therapy (NAT) using FOLFIRINOX or GNP.

Tee et al. first reported a large series of AR combined with new-generation NAT for advanced PCs [33]. In this study, 111 patients underwent pancreatectomy with AR, including any hepatic (54%), any celiac (44%), any superior mesenteric (14%) or multiple ARs (14%), with revascularization in 55% (Figure 1F). TP was performed on 20 (18%) patients. The majority of cases underwent planned AR (77%), and most of the procedures were performed post-2010 (78%). The most common indication for pancreatectomy was for PC in 87 (78%) patients. Of these patients, 65 (75%) were treated with neoadjuvant systemic chemotherapy that included FOLFIRINOX, GNP or both, with the majority (88%) also receiving sequential chemoradiation with a total dose of 50.4Gy with various radiation sensitizers. Ninety-day major morbidity (≥grade III) and mortality was 54% and 13% mainly due to post-pancreatectomy hemorrhage, postoperative pancreatic fistula or ischemia. They emphasized that a significant decrease in mortality was achieved in patients who underwent ARs post-2010 (9% compared with 29% in patients before 2010, *p* = 0.02). From the same group, Truty et al. reported a systematic classification of CAR, which included three levels according to the extent of the resection: class 1, celiac only, class 2, celiac and PHA and class 3, SMA additional to class 1 or 2 [57]. Ninety-day mortality was 10%, with a significant improvement in the last 50 consecutive cases (4%). The R0 resection rate (88%) was associated with chemoradiation (*p* = 0.004). The MST was 36.2 months, superior from the neoadjuvant chemotherapy (8.0 vs. 43.5 months). Truty et al. also reported a large series comprising 194 borderline resectable or LAPC [87]. En-bloc venous and/or arterial resection was required in 125 (65%) patients, with 94% of patients achieving R0 margins. TP was performed in 25 (13%) patients. The 90-day mortality was 6.7%. Among patients without mortality, epochally favorable survival outcomes were obtained (the median, one-year, two-year and three-year overall survivival (OS) rates were 58.8 months, 96%, 78% and 62%, respectively). They emphasized the efficacy of total neoadjuvant therapy (TNT) with favorable prognostic factors: extended duration (six cycles) of neoadjuvant chemotherapy, optimal post-chemotherapy CA19-9 response and major pathological response. Bachellier et al. reported a large AR series for PCs with excellent postoperative outcomes [22]. The most impactful point was that this study included 35 SMA resections, which was the largest ever. The overall mortality and morbidity were 5.1% and 41.5%, respectively. Preoperative radiation was not employed, assumably to secure the safety of complicated AR of the major visceral arteries. TP was performed in 18 (15%) patients. Some patients (75.4%) underwent NAT. The median, one-year, three-year and five-year OS rates after resection were 13.7 months, 59%, 13% and 12%, respectively. They identified that R0 resection (hazard ratio: 0.60, *p* = 0.01) and pathological venous invasion (hazard ratio: 1.67, *p* = 0.04) were independent prognostic factors. Loos et al. reported the largest AR series (195 patients) for LAPCs recently [23]. They compared AR with periadventitial dissection (PAD; *n* = 190), which was an optional technique to achieve R0 resection in borderline resectable or LAPCs, and revealed higher rates of postoperative pancreatic fistula (4.2% after PAD vs. 10.3% after AR; *p* = 0.022), post-pancreatectomy hemorrhage (4.7% vs. 14.9%; *p* = 0.001), ischemia (4.2% vs. 15.9%; *p* < 0.0001) and relaparotomy (12.6% vs. 26.9%; *p* = 0.001) after AR. The overall mortality rate of AR was higher than that of PAD (12.8% vs. 4.7%; *p* = 0.005). Although the mortality rate became lower and lower through the study period, AR remained more dangerous than PAD. The authors concluded even experienced pancreatic surgeons needed a learning curve of 15 ARs to safely perform the procedure. These results indicated the difficulty of AR to be disseminated globally. The median and five-year OS rates were 21.5 months and 15%, respectively, after PAD and 17.7 months and 9% after AR (*p* = 0.099). These results were attributed to more advanced stages and less incidences of NAT in the AR group.

#### *4.2. En-Bloc Arterial Resection or Arterial Divestment?*

There still remains controversy over the issue of whether we choose AR for major vessels or not, especially for the SMA. Even in highly selected patients, the SMA resection is regarded as difficult to be generalized. To balance surgical and oncological safety, the arterial divestment technique has been proposed as an alternative for SMA resection. "Divestment" means "undressing" or "circumferential dissection". The detailed technique and outcomes of arterial divestment were described in recent reports from the Heidelberg group [88,89]. The SMA was dissected using an artery-first approach through a wide Kocher maneuver, and if needed, a Cattel-Braasch maneuver was added. The authors recommend intraoperative sampling of the periadventitial tissue around the SMA, and if the cancer was positive, divestment was first attempted. Cai et al. recommended in their report that peri-adventitial dissection should be done with cold dissection using the tip of a rightangled clamp or the nonworking tip of energy devices [89]. Burn injury on the arterial wall would be a risk of postoperative aneurysm. If the dissection was difficult due to direct encasement, finally, AR was employed. To select among the three choices: divestment, AR or aborting resection before the point of no return, an artery-first approach is mandatory. The safety of the divestment technique was reported by a recent article from the same group of Heidelberg [23]. Inoue et al. [16] described the details of periadventitial dissection around the SMA, which resulted in no mortality by the use of an artery-first approach. It did not preclude postoperative recovery or adjuvant therapy if the neurogenic diarrhea was adequately controlled. However, the safe utilization of this technique has never

been generalized. Sabater et al. [90] conducted the first randomized trial to compare the oncological and surgical outcomes between artery-first PD and standard PD. The authors concluded that they found no difference either in the R0 resection rates (67.9 % vs. 77.3 %, *p* = 0.194) or in the postoperative complications (overall morbidity rate; 67.9% vs. 73.3%, *p* = 0.484) in patients undergoing artery-first PD versus standard PD. Although this trial included only resectable PCs and other periampullary malignancies, and their conclusions could not be applied directly to the management of LAPCs, this technique should be carefully applied by an expert pancreatic surgeon at a high-volume center. Another important matter is when and how we decide the approach to the SMA. Habib et al. [91] also encouraged SMA divestment for selected patients after new-generation NAT. They also indicated the usefulness of the preoperative radiological finding of circumferential SMA encasement. Halo sign, wherein the SMA was surrounded by hypodense tissue without narrowing, was potentially a candidate for resection using arterial divestment. On the other hand, string sign, wherein the SMA was surrounded by periadventitial tissue forming an irregular narrowing (like a string), was not a candidate for R0 resection, even with arterial divestment. Habib et al. and the John's Hopkins group did not regard a patient with string sign as an adequate candidate for resection, because they could not justify SMA resection due to the high morbidity and mortality. However, the radiological change after NAT did not represent a pathological regression of the tumor cells, and decision-making by the preoperative findings alone would include the risk of overdiagnosis and loss of chance for a cure. Del Chiaro also advocated intraoperative decision-making of the divestment or AR [92]. The author also recommended performing the divestment technique by the surgical team experienced in AR, because we have to prepare for unexpected arterial injury during SMA dissection, which requires complex vascular reconstruction.

On the other hand, Truty et al. [87] strongly recommended a planned en-bloc resection, even for the SMA. Their recent report still included a high mortality rate (9 out of 71 LAPC patients) after aggressive AR, but they stated that the safety of AR has become more robust recently and documented a surprisingly high R0 rate and long-term survival. Actually, the intraoperative judgement of periadventitial cancer invasion requires a test dissection, which potentially cuts into the cancer tissue. The superiority of planned en-bloc portal vein resection in obtaining R0 to unplanned venous resection after a test dissection was recently documented [93]. The en-bloc approach is exactly the principle of regional pancreatectomy suggested by Fortner et al. [7], and the reappraisal of regional en-bloc resection has been reported, such as for portal vein resection [94]. If the safety of AR is guaranteed, the same theory should be justified in SMA resection as well. For pancreatectomy with complicated AR, the efficacy of concomitant TP has been reappraised. The total removal of the pancreatic gland makes the procedure safer by eliminating the problem of pancreas fistula and its potentially fatal effect on arterial anastomosis [95,96]. This strategy, which was originally suggested at the dawn of the radical resection of PCs, has become justified after the improvement of the perioperative management of TP patients through several decades.

#### *4.3. Rationale of Total Neoadjuvant Therapy*

Another recent topic relevant to extremely radical pancreatectomy for PCs is the rationale of TNT. TNT has been advocated for LA gastrointestinal cancers, i.e., esophageal cancers [97] or rectal cancers [98,99], wherein the surgical burden of resection likely hampers prompt postoperative recovery and adequate adjuvant systemic chemotherapy. For LAPCs, due to a lack of effective regimens, TNT has long been out of the question, and the efficacy of TNT was suggested only recently. Murphy et al. reported a prospective single-arm phase II trial evaluating the efficacy of TNT using FOLFIRINOX for LAPCs with the primary endpoint of the R0 resection rate [100]. This report was the first concrete evidence of TNT for PCs. Forty-nine LAPC patients were enrolled. Eight cycles of FOLFIRINOX were administered, followed by short- or long-course chemoradiotherapy, depending on the radiological findings after FOLFIRINOX. Thirty-nine (80%) patients completed eight cycles. One patient (2%) had a radiographic complete response. Twenty-three patients

(49%) had a partial response, while 21 (45%) had a stable disease. Two patients (4%) had a progressive disease by the response evaluation criteria in solid tumours (RECIST) criteria. Thirty-four patients (69%) underwent surgical resection. Finally, 30 (61%) patients achieved R0 resection. TNT with FOLFIRINOX was feasible and provided a favorable long-term survival (median progression-free survival was 17.5 months (95% CI: 13.9–22.7), and median MST was 31.4 months (95% CI, 18.1–38.5)). For LAPCs, intensive neoadjuvant therapy has already become a consensus, and the next issue is how we can standardize the optimal contents, dose and duration of NAT. Moreover, scientifically reliable evidence for neoadjuvant therapy for PCs [101,102] is still sparse compared to adjuvant therapy [103–106] so far. Whether or not we should really omit adjuvant therapy remains unclear.

#### **5. Conclusions**

In this review, the recent development of radical pancreatectomy, including arterial resection, arterial divestment or total pancreatectomy, was discussed. Thanks to the recent improvement of chemotherapy using multiple agents, both tumor suppression and patient selection have become pragmatic. Simple resection of the HA or CA and TP has likely become a matured technique. To implement ERP including SMA resection or combined major arterial resections, the further accumulation of cases, the establishment of a standardized technique and optimal neoadjuvant therapy should be pursued.

**Author Contributions:** Conceptualization, Y.I. and Y.T.; methodology, Y.I.; writing—original draft preparation, Y.I.; writing—review and editing, A.O., Y.O., T.S. and H.I. and supervision, Y.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Clinical Benefits of Conversion Surgery for Unresectable Pancreatic Ductal Adenocarcinoma: A Single-Institution, Retrospective Analysis**

**Yuko Mataki 1,\* , Hiroshi Kurahara <sup>1</sup> , Tetsuya Idichi <sup>1</sup> , Kiyonori Tanoue <sup>1</sup> , Yuto Hozaka <sup>1</sup> , Yota Kawasaki <sup>1</sup> , Satoshi Iino <sup>1</sup> , Kosei Maemura <sup>2</sup> , Hiroyuki Shinchi <sup>3</sup> and Takao Ohtsuka <sup>1</sup>**


**Citation:** Mataki, Y.; Kurahara, H.; Idichi, T.; Tanoue, K.; Hozaka, Y.; Kawasaki, Y.; Iino, S.; Maemura, K.; Shinchi, H.; Ohtsuka, T. Clinical Benefits of Conversion Surgery for Unresectable Pancreatic Ductal Adenocarcinoma: A Single-Institution, Retrospective Analysis. *Cancers* **2021**, *13*, 1057. https://doi.org/10.3390/cancers 13051057

Academic Editor: Adam E. Frampton

Received: 1 February 2021 Accepted: 25 February 2021 Published: 2 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Simple Summary:** Pancreatic ductal adenocarcinoma (PDAC) is a dismal disorder, but conversion surgery (CS) has provided possibilities of better prognosis for unresectable (UR-)PDAC. We retrospectively investigated the clinical benefits of CS in patients with UR-PDAC. We performed CS in 20 of the 398 UR-PDAC cases between 2006 and 2019(5.1%). Compared the overall survival (OS) period among patients undergoing CS, resectable (R), borderline resectable (BR), unresectable locally advanced cancer (UR-LA), and unresectable cancer with distant metastasis (UR-M) groups, the CS group had significantly better survival than R, BR, UR-LA, and UR-M groups (73.7, 32.7, 22.7, 15.7, and 8.8 months, respectively). Moreover, multivariate analysis revealed the presence of chemoraiotherapy and partial response/complete response in the Response Evaluation Criteria in Solid Tumors (RECIST) were statistically significant prognostic factors for OS among patients undergoing CS (*p* = 0.004 and 0.03, respectively). This study highlights importance of multidisciplinary treatment including CS for patients with UR-PDAC.

**Abstract:** Background: Unresectable pancreatic ductal adenocarcinoma (UR-PDAC) has a poor prognosis. Conversion surgery is considered a promising strategy for improving the prognosis of UR-PDAC. This study aimed to investigate the clinical benefits of conversion surgery in patients with UR-PDAC. Methods: We retrospectively evaluated patients with PDAC who were referred to our department for possible surgical resection between January 2006 and December 2019. Conversion surgery was performed only in patients with UR-PDAC who could expect R0 resection. We analyzed the prognostic factors for overall survival among patients who underwent conversion surgery. Results: Overall, 638 patients with advanced pancreatic cancer were enrolled in this study. According to resectability, resectable cancer (R) was present in 180 patients, borderline resectable cancer (BR) was present in 60 patients, unresectable locally advanced cancer (UR-LA) was present in 252 patients, and unresectable cancer with distant metastasis (UR-M) was present in 146 patients. Conversion surgery was performed in 20 of the 398 UR cases (5.1%). The median period between the initial therapy and conversion surgery was 15.5 months. According to the Response Evaluation Criteria in Solid Tumors (RECIST) evaluation, the treatment response was CR in one patient, PR in 13, SD in five, and PD in one. Downstaging was pathologically determined in all cases. According to the Evans grading system, grade I was observed in four patients (20%), grade IIb was observed in seven (35%), III was observed in seven (35%), and IV was observed in two (10%). We compared the overall survival period from initial treatment among patients undergoing conversion surgery; the median overall survival durations in the conversion surgery, R, BR, UR-LA, and UR-M groups were 73.7, 32.7, 22.7, 15.7, and 8.8 months, respectively. Multivariate analysis revealed that the presence or

absence of chemoradiotherapy (CRT) and the RECIST partial response (PR)/complete response (CR) for the main tumor were statistically significant prognostic factors for overall survival among patients undergoing conversion surgery (*p* = 0.004 and 0.03, respectively). Conclusion: In UR-PDAC, it is important to perform multidisciplinary treatment, including CRT with conversion surgery.

**Keywords:** unresectable pancreatic ductal adenocarcinoma; conversion surgery; chemoradiotherapy

#### **1. Introduction**

Pancreatic ductal adenocarcinoma (PDAC) is a very dismal disease with a poor prognosis among malignant tumors in Japan [1]. Even in the United States, PDAC is the fourth leading cause of cancer-related deaths [2]. This disease has a 5-year survival rate of approximately 10%. Surgical resection is the only potentially curative therapy, but only 10–20% of patients with resectable PDAC are classified at the time of the initial diagnosis [3]. When we select the methodology of treatment for a patient with PDAC, the appropriate treatment is likely to decided based on the resectable classification rather than the stage classification. The National Comprehensive Cancer Network (NCCN) has proposed a resectable classification for PDAC and recommended optimal therapy based on each resectability [4]. Among them, pancreatic cancer patients with unresectable locally advanced cancer (UR-LA) and unresectable cancer with metastasis (UR-M) are recommended to receive chemotherapy if their performance status is good. Combination chemotherapy with fluorouracil, leucovorin, irinotecan, and oxaliplatin (FOLFIRINOX) or modified (mFOLFIRINOX), chemoradiation, gemcitabine plus nab-paclitaxel (PTX), and/or chemoradiation are listed as candidate regimens for chemotherapy. The Japanese Pancreas Society designed new and definite resectability criteria in 2019 based on the NCCN guideline [5]. Regarding resectable advanced pancreatic cancer, a prospective randomized trial in Japan compared surgical resection and chemoradiotherapy (CRT) after laparotomy was used to identify the localized pancreatic cancer invading the surrounding vasculature; this study showed that patients who underwent surgical resection had significantly longer survival than those treated with CRT alone [6].

In recent reports, multidisciplinary treatments, including surgical resection and chemotherapy or chemoradiation therapy, have improved the survival of patients with PDAC [7–9]. As for patients with resectable (R) PDAC, adjuvant chemotherapy has become standard even after curative resection [10,11]. Moreover, many studies have reported the effectiveness of neoadjuvant chemotherapy. In recent years, patients with UR-PDAC have been able to achieve treatment effects to be converted to surgical resection. This surgical strategy is called "conversion surgery". Therefore, several reports about conversion surgery have been published in selected patients with initially UR-PDAC, and the prognostic effect of conversion surgery has been reported [12–22]. However, it is unclear whether conversion surgery with a high treatment effect is truly advantageous in patients with UR-PDAC.

In this study, we aimed to clarify the clinical benefits and important factors of conversion surgery for patients with UR-PDAC who underwent CRT.

#### **2. Methods**

#### *2.1. Ethics Statements*

All research performed in studies involving human participants were according to discipline of the institutional research committee and with the 1964 Declaration of Helsinki. Written informed consent was obtained from all study participants.

#### *2.2. Study Design and Patient Population*

This retrospective study was performed using data from a prospective database. All patients with PDAC treated at the Department of Digestive Surgery, Breast and Thyroid Surgery, Kagoshima University between January 2006 and December 2019 were enrolled in

our study. PDAC was diagnosed by cytology or pathology through endoscopic retrograde cholangiopancreatography or endoscopic ultrasound-guided fine-needle aspiration. PDAC progression was diagnosed using multidetector-row computed tomography, ethoxybenzylmagnetic resonance imaging, and fluorine-18-2-deoxy-D-glucose positron emission tomography. All patients were divided into resectable cancer (R), borderline resectable cancer (BR), UR-LA, and UR-M, which was defined according to the NCCN Clinical Practice Guidelines version 2 from 2020 [4].

#### *2.3. Therapy Method*

The gemcitabine and S-1 (GS) regimen is as follows, that the daily dosage of S-1 chemotherapy was 60 mg/m<sup>2</sup> combined with 1000 mg/m<sup>2</sup> gemcitabine. Patients received gemcitabine intravenously on days 8 and 15 of a 21-day cycle and oral S-1, twice daily on days 1–14. The GEM plus nab-PTX regimen consisted of gemcitabine 1000 mg/m<sup>2</sup> and nab-PTX 125 mg/m<sup>2</sup> administered on days 1, 8, and 15, every 28 days. The modified FOLFIRINOX (m-FOLFIRINOX) regimen used in our study was the same as that of a Japanese phase II study and consisted of oxaliplatin at a dose of 85 mg/m<sup>2</sup> , administered as a 2-h intravenous infusion, immediately followed by leucovorin at a dose of 200 mg/m<sup>2</sup> , administered as a 2-h intravenous infusion, with the addition, after 30 min of irinotecan at a dose of 180 mg/m<sup>2</sup> , administered as a 90-min intravenous infusion through a bypass line. This treatment was immediately followed by fluorouracil at a dose of 400 mg/m<sup>2</sup> administered via intravenous bolus, followed by a continuous intravenous infusion of 2400 mg/m<sup>2</sup> over a 46-h period, every 2 weeks. CRT regimens included hyper-fractionated accelerated radiotherapy administered with S-1 at 80 mg/m<sup>2</sup> for the first 21 days. A total of 50–58 Gy was administered in 40 fractions over 4 weeks. At 1 month after CRT completion, S-1 was administered for 2 weeks, followed by a 2 weeks period.

The GEM plus nab-PTX and FOLFIRINOX regimens were approved in Japan for the treatment of UR-PC in December 2014, and these regimens have been utilized since then.

#### *2.4. Conversion Surgery*

We discussed and decided the indication for conversion surgery for individual UR-PDAC patients at a multidisciplinary conference of pancreatic surgeons, medical oncologists, radiologists, and clinical pathologists. Among UR-PDAC patients who responded to various therapies, conversion surgery was permitted for only those who met the following conditions: patients showing adequate reduction of the main tumor, enabling complete removal inclusive of the major vessels and metastatic site, those with at least several months of local control, those with no metastasis, or those with controllable metastasis by surgical resection.

#### *2.5. Adjuvant Chemotherapy*

We basically consider adjuvant therapy after conversion surgery if the patient is in good condition. Furthermore, if new lesions are operable after conversion surgery, we consider additional resection for the lesions.

#### *2.6. Assessment*

The clinical treatment effect was assessed using RECIST version 1.1 [23], and the histologic assessment of the extent of response was evaluated using the Evans grading system [24]. R0 was defined as pathologically margin free in the resected specimen.

The Clavien–Dindo classification was used to evaluate postoperative complications [25]. Mortality was defined as death within 90 days after surgery.

#### *2.7. Statistical Analysis*

Associations between different categorical variables were properly assessed using the chi-square or Fisher exact test. Survival curves were estimated using the Kaplan–Meier method and analyzed using the log-rank test. Overall survival was calculated as the

interval between initial treatment and death due to any cause (including death from other diseases), whereas progression-free survival was calculated as the interval between initial treatment and disease progression. The univariate Cox proportional hazard models were used to estimate the independent significant factors for the overall survival of patients with UR-PDAC. Relative risks were expressed as adjusted hazard ratios and corresponding 95% confidence intervals. Valuables with a *p*-value of < 0.05 on univariate analysis were calculated into the multivariate model. A *p*-value of < 0.05 was considered to be statistically significant. All statistical analyses were performed using SigmaPlot, version 12.5 for Windows (HULINKS, Inc., Tokyo, Japan).

#### **3. Results**

#### *3.1. Patient Characteristics*

During the above observation period, 638 patients were included. R was present in 180 patients, BR was present in 60, UR-LA was present in 252, and UR-M was present in 146. Overall, 398 consecutive patients with unresected pancreatic cancer were studied in terms of their unresectable status. Conversion surgery was performed in 20 patients (20/398, 5.0%; UR-LA, *n* = 9; UR-M, *n* = 11). The clinical characteristics of patients with and without conversion surgery are compared in Table 1.



M, male; F, female; CS, conversion surgery; Ph, pancreatic head; Pb, pancreatic body; Pt, pancreatic tail; UR-LA, unresectable locally advanced cancer; UR-M, unresectable cancer with metastasis. T, N, and M were classified according to the Tumor-Node-Metastasis (TNM) classification.

There were no statistically significant differences in age, sex, the tumor location (pancreatic head versus (vs.) pancreatic body and tail), unresectability status (UR-LA vs. UR-M), serum carcinoembryonic antigen (CEA) level, cancer antigen 19-9 (CA19-9) level, tumor size, and T-, N-, and M-categories according to the Tumor-Node-Metastasis (TNM) classification before treatment between patients with and without conversion surgery.

The clinical characteristics of the 20 patients who underwent conversion surgery are shown in Figures 1 and 2 and Tables 2 and 3.

**Figure 1.** Clinical characteristics of the 20 patients who underwent conversion surgery including therapy before operation. Abbreviations: UR-LA, unresectable locally advanced cancer; UR-M, unresectable cancer with metastasis; M, male; F, female; Ph, pancreatic head; Pb, pancreatic body; Pt, pancreatic tail; LN16, para-aortic lymph node metastasis; PER, peritoneal disseminations; HEP, hepatic metastases; CRT, chemoradiotherapy; GEM, gemcitabine; S-1, oral administration of S-1; nabPXL, nab-paclitaxel; mFOLFIRINOX, modified FOLFIRINOX (combination chemotherapy with fluorouracil, leucovorin, irinotecan, and oxaliplatin).

**Figure 2.** Clinical characteristics of the 20 patients who underwent conversion surgery including therapy after operation. Abbreviations: pleural, pleural disseminations; LM, liver metastasis; lung mets, lung metastasis; LR, local recurrence; peritoneal, peritoneal disseminations; directing arrow, recurrent or metastatic period; Ope, additional operation; +, period of death.




**Table 3.** Clinical characteristics of 20 patients who underwent conversion surgery. Each stage was classified according to the TNM classification.

UR-LA, unresectable locally advanced cancer; UR-M, unresectable cancer with metastasis; before, before initial therapy; after, after various therapy; *p*-stage, pathological stage; tub1, tubular adenocarcinoma with high differentiation; tub2, tubular adenocarcinoma with moderate differentiation; por, poorly differentiated adenocarcinoma; muc, mucinous adenocarcinoma; NEN, neuroendocrine neoplasm; a few, a few neoplasms; Evans, Evans classification.

Among the 11 patients with UR-M, the type of metastasis at the time of initial diagnosis was hepatic metastasis in five patients, peritoneal dissemination in three, and para-aortic lymph node metastasis in three. Histological confirmation of the metastatic lesions was made in all patients with peritoneal disseminations and para-aortic lymph node metastasis, but this was not the case in three of five patients with liver metastasis.

The periods between initial treatment and conversion surgery were a median of 10.8 (range, 5.1–19.9) months in patients with UR-LA and median of 13.3 (range, 4.4–75.9) months in those with UR-M. There were no significant differences in the period between the groups.

The methods of therapy until conversion surgery varied, but CRT was performed in all patients with UR-LA. Among them, CRT was selected as the initial therapy in all patients except one (case 2, Figure 1), and CRT was performed after gemcitabine plus nab-PTX therapy as induction chemotherapy in the remaining case (case 2, Figure 1).

Chemotherapy of various types was performed in patients with UR-M, and additional CRT was carried out to perform local control in five of these cases. According to the RECIST evaluation, the treatment response after various therapies was complete response (CR) in one patient, partial response (PR) in 13, stable disease (SD) in five, and progressive disease (PD) in one.

The CA19-9 level at the time of conversion surgery was significantly lower than that at the time of initial diagnosis (median, 16.4 U/mL (range, 0.6–99) vs. median, 156 U/mL (range 0.6–1985), *p =* 0.0065), but the CEA level was not significantly different at the time between the initial diagnosis and conversion surgery (*p =* 0.418). The size of the main tumor at the time of conversion surgery was significantly smaller than that at the time of diagnosis (median, 13 mm (range, 0–50) vs. 30 mm (range, 18–50), *p <* 0.001). In cases 6, 17, and 19, the CEA level, CA19-9 level, and size of the main tumor after each therapy were reduced compared with those in the period before initial therapy, respectively (Table 2).

As for the operative method, substomach-preserving pancreaticoduodenectomy was performed in 13 patients, including concomitant portal vein resection (PVR) in six patients and hepatic resection (HPR) in one patient. Distal pancreatectomy and total pancreatectomy with PVR and HPR were performed in six patients and one patient, respectively (Table 2). The median operative time and volume of blood loss were 575 (range, 352–919) min and 1150 (range, 90–3600) mL, respectively. The median postoperative stay was 13 (range, 8–50) days. Major complications in the postoperative period (Clavien–Dindo classification class IIIa) occurred in three patients with chylous ascites, interstitial pneumonia, and intraabdominal abscess (cases 1, 3, and 14, respectively; Table 2); those complications caused patients' postoperative hospital stay to exceed 1 month. No perioperative mortality was observed. R0 resection of the main tumor was performed in all patients. Compared with before and after various therapies, clinical stage was downstaged in 13 cases, and the pathological stage was downstaged in 19 cases. Pathologically, tubular adenocarcinoma with high differentiation and tubular adenocarcinoma with moderate differentiation were the pathological forms of the residual tumor in six and eight cases, respectively. According to the Evans grading system, grade I was observed in four patients (20%), grade IIb in seven (35%), III in seven (35%), and IV in two (10%). A pathological CR was observed in two patients (cases 10 and 16, Table 2). The pathological assessment was difficult to perform in one case because of a few residual cancer cells (case 20, Table 2). In two patients who underwent hepatic resection combined with pancreatectomy, one patient had residual cancer cells in the resected liver.

#### *3.2. Adjuvant Chemotherapy*

Only eight patients (40%) received postoperative adjuvant chemotherapy, and S-1 was administered to all those patients. The reasons why 12 patients did not receive adjuvant chemotherapy were poorer performance status, severe perioperative complications, elderly age, and patient unwillingness. Recurrence occurred in 14 patients (70%) who underwent conversion surgery and received various therapies. The recurrence type varied: dissemination occurred in four patients, local recurrence, including metachronous double cancer, in three, liver metastasis in three, lung metastasis in three, and lymph node metastasis in one. Two patients with lung metastasis and one with metachronous double cancer of the remnant pancreas underwent additional resection of the metastatic lesion at 41, 78, and 97 months after conversion surgery, respectively (cases 3, 6 and 7; Figure 2).

We compared overall survival from the period of initial treatment between patients who underwent conversion surgery with UR-LA and those with UR-M and found no significant difference between these patient groups (median not reached vs. 52 months, *p =* 0.20) (Figure 3).

Moreover, we compared overall survival from the period from initial treatment between the conversion surgery, R, BR, UR-LA, and UR-M groups according to the NCCN guidelines. The median overall survival durations in the conversion surgery, R, BR, UR-LA, and UR-M groups were 73.7, 32.7, 22.7, 15.7, and 8.8 months, respectively (Figure 4). There were significant differences in the survival period between each group (*p*-value not shown).

By the univariate and multivariate logistic regression analyses, we researched the prognostic factors for overall survival among patients who underwent conversion surgery. Table 4 shows that the presence of CRT and RECIST PR/CR for the main tumor decreased the risk of death relative to the absence of CRT and RECIST SD/PD for the main tumor

(*p =* 0.002 and 0.015, respectively). Multivariate analysis revealed that the presence of CRT and RECIST PR/CR for the main tumor was a statistically significant prognostic factor for overall survival among patients undergoing conversion surgery (*p =* 0.004 and 0.03, respectively).

**Figure 3.** Comparison of the overall survival curves between patients with unresectable locally advanced cancer (dashed line, *n* = 9) and unresectable cancer with metastasis (solid line, *n* = 11) who underwent conversion surgery. There is no significant difference in overall survival between the groups (*p =* 0.20).

**Figure 4.** Comparison of the overall survival curves between patients with pancreatic ductal adenocarcinoma judged as R (*n* = 180), BR (*n* = 60), UR-LA (*n* = 243), or UR-M, (*n* = 135) according to the NCCN guideline, who underwent conversion surgery (*n* = 20). Abbreviations: R, resectable cancer; BR, borderline resectable cancer; UR-LA, unresectable locally advanced cancer; UR-M, unresectable cancer with metastasis; NCCN, National Comprehensive Cancer Network.


**Table 4.** Predictive factors for the overall survival of patients with unresectable pancreatic cancer (univariate and multivariate logistic regression analyses).

chemoradiotherapy; PD, progressive disease; SD, stable disease; PR, partial response; CR, complete response; LN mets, lymph node metastasis.

#### **4. Discussion**

The reports on conversion surgery for locally advanced pancreatic cancer have been increasing since the 2010s [12–22]. According to the recent NCCN guidelines (version 1.2021) [4], surgical resection of locally advanced cancer is a subsequent therapy option for patients with good performance status and no disease progression after first-line therapy. The present retrospective study found that conversion surgery for UR-PDAC after various therapy is relatively safe and feasible, because of no mortality and lower severe morbidity (C-D classification ≥ IIIa was 15% rate), and it may help improve long-term prognosis.

In the current study, the resection rate was only 5.0% (UR-LA, 3.6%; UR-M, 7.5%). However, this value was very low compared with that in previous reports; the reason for the resection rate might depend on policy whether we actively accepted conversion surgery when multidisciplinary treatment was performed for patients with UR-PDAC [16]. Many studies have focused on the effectiveness of conversion surgery, and although patients with UR-PDAC were not candidates for surgery, conversion surgery was performed for only an extremely limited number of patients who were super-responders to palliative therapy. Hackert et al. reported that among 575 patients with locally advanced PDAC receiving neoadjuvant therapy, 292 (50.8%) underwent pancreatic resection, which was a higher resection rate than that in our study [7]. However, the median survival time (MST) after adjuvant surgery was only 15.3 months. In the current study, the median overall survival was 73.7 months, which is better than that in previous reports [13,16,18]. The MST of patients with UR-LA treated with no resection was reported to be 15–17 months, which was similar to that in the present study. The median overall survival durations in PDAC patients with UR-LA and UR-M were 15.7 and 8.8 months, respectively. On the other hand, Lee et al. reported an excellent outcome, showing that 15 of 64 (23.4%) PDAC patients with UR-LA who underwent conversion surgery had an overall survival exceeding 40 months, although they did not clearly describe the indications for surgery [26]. Strict criteria similar to those in our study may lead to lower resectability but longer overall survival as a result of patient selection, so we may need to expand our surgical indication.

Conversion surgery has provided a surprising outcome for patients with UR-PDAC. In particular, the prognosis of patients with UR-PDAC who underwent conversion surgery was significantly better than that of patients with resectable PDAC. Several authors have previously reported the usefulness of conversion surgery in such patients, as well as favorable results on prognosis [19,27–30]. In the present study, the prognosis of PDAC patients with UR-LA and UR-M who underwent conversion surgery was not significantly different. Among the patients with UR-PDAC who underwent conversion surgery, we identified that RECIST CR/PR after the various treatments was an important prognostic factor by multivariate analysis. It may be important to perform multidisciplinary treatment aimed at tumor shrinkage and possible conversion surgery in order to obtain a better prognosis. However, recurrence and metastasis after conversion surgery were observed in 70% of patients, especially early recurrence within one year after surgery in 30% of patients. Uesaka et al. demonstrated the non-inferiority of S-1 to gemcitabine as adjuvant chemotherapy for pancreatic cancer in terms of overall survival in a randomized trial [10]. Chad et al. reported that adjuvant therapy was not required for all patients with localized pancreatic cancer who had received neoadjuvant therapy; the benefit of adjuvant therapy was limited to those with node-positive disease [31]. In the current study, adjuvant chemotherapy after conversion surgery was performed in only eight (40%) of 20 patients. In other reports, early recurrence after conversion surgery occurred in approximately 30% of cases [32,33]. The conversion surgery tended to have a long operative time and high bleeding volume in the present study, so it was considered a major invasive surgery. Therefore, it may be important to reduce morbidity in surgery as much as possible and to create an environment in which adjuvant chemotherapy can be easily performed. The early recurrence rate should be decreased as much as possible in patients undergoing conversion surgery [34]. Moreover, in the present study, during the follow-up after conversion surgery, additional surgical resection was performed in one patient with double cancer in the remnant pancreas and

two patients with single lung metastasis. The surgical resection for multicentric cancer in the remnant pancreas and solitary lung metastasis may have possibilities to provide long-term survival. [35,36].

Satoi et al. previously performed multicenter joint research that focused on conversion surgery of initially unresectable PDAC [13]. This study included 58 PDAC patients (41 with UR-LA and 17 with UR-M). The MST in the conversion surgery group was significantly better than that in the control group (39.7 months vs. 20.8 months, *p* < 0.0001). The optimal timing of adjuvant surgery was 240 days after the initial treatment. Moreover, a multivariate analysis showed that for the adjuvant surgery group, significant favorable factors for overall survival included the dosage above a certain level of gemcitabine, a decrease in tumor markers until conversion surgery, and PR/CR evaluation by RECIST. Michelakos et al. reported that in resected patients with BR/locally advanced PDAC treated with FOLFIRINOX, a preoperative CA19-9 level >100 U/mL and >8 months between diagnosis and surgery predicted a shorter postoperative DFS [33]. Therefore, the optimal period between the initial diagnosis and conversion surgery is controversial.

As for the pathological examination, the Evans grade, which reflects the extent of tumor degeneration or necrosis after chemo(radio)therapy, has been extensively identified as a prognostic factor [19,37–39]. Chatterjee et al. reported that among the 223 patients with resectable PDAC who received neoadjuvant chemoradiation and had pancreaticoduodenectomy, pCR (grade IV, 2.7%) or minimal residual tumor (grade III, 16.1%) in posttreatment specimens of pancreaticoduodenectomy correlated with better survival [40]. Histologically, grades IIb, III, and IV, i.e., destroyed areas over 50%, were observed in the main tumor in 16 of the 20 patients who underwent conversion surgery in the present study. Moreover, the residual tumor tended to contain a wide range of highly differentiated cells, so this result suggests that cells with poorly differentiation may be highly effective for chemo(radio-)therapy.

In the present study of patients with UR-PDAC who underwent conversion surgery, we identified important prognostic factors from initial treatment in multivariate analysis: the use of chemoradiotherapy and RECIST CR/PR after various treatments. Of the treatment for UR-PDAC, radiation therapy combined with systemic chemotherapy is one of the recommended therapies in the NCCN [4] and Japanese guidelines [5]. In our institution, we reported the usefulness of the combination therapy with S-1 and radiation in patients with UR-PDAC; it is a well-tolerated regimen that can be recommended as an effective treatment in prospective phase II trials, and it showed favorable survival with a median survival time of 14.3 months [41]. Jang et al. reported on the benefit of neoadjuvant treatment in patients with BR-PDAC, and gemcitabine-based neoadjuvant chemoradiation provided oncological benefits compared to upfront surgery by the prospective randomized controlled trial [42]. A recent LAP-07 trial comparing chemotherapy and CRT for locally advanced PDAC failed to show any survival benefit of CRT [43]. However, CRT was associated with a decreased local progression rate and no increase in grade 3 or 4 toxicities. This result may indicate that CRT has potential as a more useful method in conversion surgery.

In the present study, we identified no significant difference in overall survival between patients who underwent conversion surgery with UR-LA and UR-M. Yanagimoto et al. reported no significant differences between the two groups, which is similar to our study finding [18]. There are few reports of surgical resection of pancreatic resection with synchronous metastases. Wright et al. reported 23 cases of surgical resection of stage IV pancreatic cancer after a favorable response to systematic chemotherapy [27]. The sites of metastasis included the liver (*n* = 16), lung (*n* = 6), and peritoneum (*n* = 2). The treated patients with stage IV disease were 1147 cases in all, so the resection rate was only 2.0%. The MST from the initial diagnosis was 34.1 months. Frigerio et al. reported that 24 (4.5%) of 535 patients diagnosed as pancreatic cancer with liver metastasis underwent surgical resection of the primary site and hepatic resection [44]. The MST after diagnosis in the study was 56 months. The limited number of patients with distant metastasis (super-responders) accounted for less than 5% of PDAC patients with UR-M. The effectiveness of conversion

surgery for metastatic PDAC remains controversial; therefore, it is indispensable to identify powerful surrogate markers for predicting long survival after conversion surgery.

This research has some limitations. First, this was a single-institute, retrospective analysis with a limited number of cases. Second, the necessity of conversion surgery is debatable, even though CRT is successful. Patients receiving chemotherapy generally develop a tolerance to it, and surgical resection may be the most useful means of local control; however, there is no evidence to support this. In Japan, PREP-04 (UMIN000017793), which is a multi-institutional prospective observational study to investigate the effects of conversion surgery on patients with initial UR-PDAC, is ongoing. Given that only selected patients who responded favorably to non-surgical treatment were targeted among all patients with UR-PDAC, a selection bias existed. Conversion rates vary among reports on conversion surgery for patients with UR-PDAC. Nitshe et al. reported a conversion rate of 28.6% for FOLFIRINOX [29], while Hackert et al. reported that it was 50.8% [7]. Currently, FOLFIRINOX or gemcitabine plus nab-PTX is recommended for patients with UR-PDAC. It is still unclear which anticancer drug is optimal for conversion surgery. Moreover, the usefulness of adjuvant chemotherapy for resectable PDAC has been demonstrated in the JASPAC01 study, but the usefulness of adjuvant chemotherapy in conversion surgery has not yet been proven [10]. As for radiotherapy, it is controversial whether radiotherapy should be administered. Compared to resectable PDAC, the extent of surgical invasion for UR-PDAC is larger, and it is necessary to consider patients' physical status, the drug used as adjuvant chemotherapy, and the administration period. Furthermore, the most relevant limitation of this study is represented by the lack of biological predictive markers that could support the selection of PDAC patients suitable for conversion surgery.

#### **5. Conclusions**

In conclusion, it is important to perform multidisciplinary treatment, including CRT with conversion surgery in patients with UR-PADC. However, many questions remain unsolved regarding the necessity of the conversion surgery including CRT, the optimal regimen, the duration of preoperative therapy, and criteria for surgical therapy. It is essential to perform prospective studies to resolve the various problems.

**Author Contributions:** Conceptualization, Y.M. and H.K.; methodology, K.T.; software, Y.H.; validation, T.I., K.T. and Y.H.; formal analysis, Y.K. and K.M.; investigation, H.K.; data curation, S.I.; writing—original draft preparation, Y.M; supervision, T.O.; project administration, H.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper.

**Data Availability Statement:** Not applicable

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


*Article*

## **Benefits of Conversion Surgery after Multimodal Treatment for Unresectable Pancreatic Ductal Adenocarcinoma**

**Hiroaki Yanagimoto <sup>1</sup> , Sohei Satoi 1,\*, Tomohisa Yamamoto <sup>1</sup> , So Yamaki <sup>1</sup> , Satoshi Hirooka <sup>1</sup> , Masaya Kotsuka <sup>1</sup> , Hironori Ryota <sup>1</sup> , Mitsuaki Ishida <sup>2</sup> , Yoichi Matsui <sup>1</sup> and Mitsugu Sekimoto <sup>1</sup>**


Received: 26 April 2020; Accepted: 28 May 2020; Published: 31 May 2020

**Abstract:** Background: Traditionally, the treatment options for unresectable locally advanced (UR-LA) and metastatic (UR-M) pancreatic ductal adenocarcinoma (PDAC) are palliative chemotherapy or chemoradiotherapy. The benefits of surgery for such patients remains unknown. The present study investigated clinical outcomes of patients undergoing conversion surgery (CS) after chemo(radiation)therapy for initially UR-PDAC. Methods: We recruited patients with UR-PDAC who underwent chemo(radiation)therapy for initially UR-PDAC between April 2006 and September 2017. We analyzed resectability of CS, predictive parameters for overall survival, and early recurrence (within six months). Results: A total of 468 patients (108 with UR-LA and 360 with UR-M PDAC) were enrolled in this study, of whom, 17 (15.7%) with UR-LA and 15 (4.2%) with UR-M underwent CS. The median survival time (MST) and five-year survival of patients who underwent CS was 37.2 months and 34%, respectively; significantly better than non-resected patients (nine months and 1%, respectively, *p* < 0.0001). MST did not differ according to UR-LA or UR-M (50.5 vs. 29.0 months, respectively, *p* = 0.53). Early recurrence after CS occurred in eight patients (18.8%). Lymph node metastasis, positive washing cytology, large tumor size (>35 mm), and lack of postoperative adjuvant chemotherapy were statistically significant predictive factors for early recurrence. Moreover, the site of pancreatic lesion and administration of postoperative adjuvant chemotherapy were statistically significant prognostic factors for overall survival in the patients undergoing CS. Conclusion: Conversion surgery offers benefits in terms of increase survival for initially UR-PDAC for patients who responded favorably to chemo(radiation)therapy when combined with postoperative adjuvant chemotherapy.

**Keywords:** unresectable pancreatic ductal adenocarcinoma; conversion surgery; early recurrence

#### **1. Introduction**

In the present-day situation, successful treatment of pancreatic ductal adenocarcinoma (PDAC) remains a therapeutic challenge, and the prognosis is generally poor [1]. Approximately 70% of patients with PDAC are not eligible for surgery, due to locally advanced or metastatic disease at the time of diagnosis [2]. Current guidelines of the National Comprehensive Cancer Network (NCCN) recommend nab-paclitaxel combined with gemcitabine (GnP) or FOLFIRINOX regimens as standard treatments for unresectable (UR) PDAC [3,4]. However, the median survival time (MST) for UR-PDAC remains low (9.2–13.5 months) [5–7]. The result of remarkable therapeutic response may occasionally become an

indication for conversion surgery (CS) [8,9], which is defined as additional surgery for patients with UR-PDAC who responded favorably to multimodal treatment. However, the incidence and clinical effects are unknown at present. In the present study, we evaluated the clinical outcomes of CS after chemo(radiation)therapy for UR-PDAC, predictive parameters for early recurrence (within six months after CS) and prognostic parameters for overall survival (OS).

#### **2. Patients and Methods**

#### *2.1. Study Population*

This retrospective study was conducted by using data from a prospective database. We recruited all consecutive patients undergoing chemotherapy or chemoradiotherapy for UR-PDAC who were to the Department of Surgery, Kansai Medical University, for any treatment between April 2006 and September 2017. All patients were diagnosed with PDAC by cytology or pathology through endoscopic retrograde cholangiopancreatography or endoscopic ultrasound-guided fine-needle aspiration. We have previously reported the details of multidetector-raw computed tomography (MDCT) imaging for the diagnosis of PDAC and to rule out distant metastasis, as well as staging laparoscopy techniques [10,11]. Moreover, multimodal image findings such as contrast-enhanced ultrasonography (CE-US), gadoxetic acid–enhanced magnetic resonance imaging (EOB-MRI), and positron emission tomography (PET) were considered, and we certainly confirmed that all patients had UR-PC initially, according to the National Comprehensive Cancer Network (NCCN) guideline version 2.2017 [3,4].

Ethical approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Informed consent: Written informed consent was obtained from all study participants.

#### *2.2. Data Collection*

The following data were collected: clinicopathological characteristics, type of chemotherapy or chemoradiotherapy, frequency of CS, rates of peri-operative morbidity and mortality, predictive parameters for early recurrence (defined as within 6 months after CS), and prognostic parameters for OS.

#### *2.3. Statistical Analysis*

Data are presented as median (range). Continuous or categorical variables were compared by using the Mann–Whitney U, chi-square, or Fisher's exact tests as appropriate. The OS and recurrence-free survival curves were estimated by using the Kaplan–Meier method and compared by using the log-rank test. Predictive factors identified by the univariate analysis were further examined by multivariate logistic regression analysis, to determine significant factors for OS and early recurrence among patients undergoing CS. The hazard ratio and 95% confidence intervals were calculated for all estimates. A two-tailed *p*-value of <0.05 was considered to be statistically significant. Calculations were performed by using JMP software, version 10 (SAS Inc., Cary, NC, USA).

#### **3. Results**

#### *3.1. Patient Characteristics*

Between April 2006 and September 2017, a total of 758 patients received treatment at our department; 290 of those patients underwent surgical resection. The remaining 468 patients with unresectable (UR) PDAC were finally enrolled in this study. Diagnoses were confirmed by using MDCT for 189 patients (40.4%) with unresectable locally advanced (UR-LA) PDAC and 279 (59.6%) with unresectable metastatic (UR-M) PDAC. We performed staging laparoscopy for 133 patients (28.4%) and palliative gastrojejunostomy for 20 patients (4.3%) with radiologically defined locally advanced disease. Positive peritoneal lavage cytology was identified in 30 patients (6.4%), peritoneal dissemination in 25 (5.3%), liver metastasis in 20 (4.3%), and other metastases in six (1.3%). In total, we treated 108 patients (23%) with UR-LA and 360 patients (77%) with UR-M (Figure 1).

**Figure 1.** Study flow diagram. Diagnoses were made, using multidetector-raw computed tomography. In total, 189 patients were diagnosed with unresectable locally advanced pancreatic ductal adenocarcinoma (UR-LA PDAC), and 279 patients were diagnosed with unresectable metastatic (UR-M) PDAC. We performed staging laparoscopy for 133 patients with radiologically defined locally advanced disease. We finally enrolled 108 patients with UR-LA PDAC and 360 patients with UR-M PDAC in the present study. Abbreviations: KMU, Kansai Medical University; MDCT, multidetector-raw computed tomography; UR-LA, unresectable locally advanced; UR-M, unresectable metastatic.

Baseline characteristics of the study population and regimens that were selected as first-line treatment are listed in Table 1. The most frequently used regimen was gemcitabine (GEM), followed by GEM combined with S-1 and GEM combined with nab-paclitaxel.




**Table 1.** *Cont.*

UR-LA: unresectable locally advanced pancreatic cancer, UR-M: metastatic pancreatic cancer, PS: performance status, GEM: gemcitabine, GS: S-1 combined with gemcitabine, GnP: nab-paclitaxel combined with gemcitabine, PTX: paclitaxel.

The standard treatment for advanced pancreatic cancer has changed to gemcitabine since 2001, FOLFIRINOX since 2010, and GnP since 2013 in Japan. Moreover, gemcitabine combined with S-1 was often used as a treatment option. There was liver metastasis in 193 patients, peritoneal metastasis in 123 patients, and LA in 108 patients, respectively. Standardized regimen of chemotherapy in each time has been used in patients with UR-M PDAC. Patients with peritoneal metastasis were treated with S-1 + intravenous and intraperitoneal paclitaxel [12]. Moreover, we have implemented additional radiation therapy in UR-LA patients who still had the low-density area around celiac artery or superior mesenteric artery just before the planned conversion surgery for expecting the margin-negative resection. Positive peritoneal washing cytology was not defined as M1 at that time. Therefore, chemoradiation therapy was implemented for UR-LA with positive cytology.

#### *3.2. Best Response After First-Line Treatment*

Radiographic partial responses (PR) according to Response Evaluation Criteria in Solid Tumors (RECIST) criteria were observed in 45 patients (42%) with UR-LA and 86 (24%) with UR-M. Stable disease (SD) was observed in 38 patients (35%) with UR-LA and 119 (34%) with UR-M, and disease progression observed in 25 patients (23%) and 155 (42%), respectively. Disease control was achieved in 83 patients (77%) with UR-LA and 205 (58%) with UR-M. Furthermore, patients who could maintain PR or SD for more than eight months were shown in 44 patients (40.7%) with UR-LA and in 85 patients (23.6%) with UR-M, respectively.

#### *3.3. Conversion Surgery*

The major eligibility criteria for surgical exploration were as follows: clinical response (PR/CR) on CT imaging, reduction of tumor markers, fine performance status with patient's willingness for surgery, and an interval of at least eight months since initial treatment [13]. In patients with peritoneal metastasis, disappearance of occult distant organ metastasis was confirmed by second-look staging laparoscopy in the context of the above criteria. In patients with liver metastasis, a maximum of three occult metastases on the liver surface were resected. In cases where tumor extension to the major vessels with attachment was observed, these patients were indicated for resection. Clinical staging and surgical exploration were re-evaluated at multidisciplinary team meetings.

During the study period, 36 patients were planned to undergo CS, and four underwent exploratory laparotomy for occult distant organ metastasis. Finally, CS was performed on 17 patients (15.7%) with UR-LA and 15 (4.2%) with UR-M. Some reasons were raised in 99 patients who had PR but did not undergo conversion surgery due to still UR-LA status on CT imaging and poor performance status. We performed subtotal stomach-preserving pancreaticoduodenectomy for 13 patients (40.6%), distal pancreatectomy for 11 (34.4%), total pancreatectomy for four (12.5%), and distal pancreatectomy with en-bloc celiac axis resection (DP-CAR) on four patients (12.5%) (Table 2). Concomitant CHA resection was done in four patients (12.5%), and concomitant portal vein resection was in 15 patients (46.9%). R0 resection was achieved in 29 patients (90.6%). The median operative time for the total study population was 441 (range 223–866) min, and the median intraoperative blood loss was 1250

(range 207–6301) mL. Although the complication of Clavien–Dindo classification ≥IIIa [14] was reported for eight patients (25.0%), there was no mortality. The median postoperative hospital stay was 14 (range 7–116) days. Histopathologically, Evans grade ≥III was noted in nine patients (28.1%), one of whom exhibited pathological complete response (pCR). The 23 patients (71.9%) received postoperative adjuvant chemotherapy; S-1 was administered to 13 patients (40.6%), GEM to three (9.4%), GEM plus S-1 to one (3.1%), and intraperitoneal infusion and intravenous administration of paclitaxel combined with S-1 to six (18.8%). Twenty-two patients (68.8%) completed adjuvant chemotherapy. The nine patients (28.1%) did not receive postoperative adjuvant chemotherapy, because of our policy of non-adjuvant chemotherapy in the first four patients, patient's willingness (*n* = 3), or insufficient nutritional condition (*n* = 2).


**Table 2.** Patient characteristics of conversion surgery.

Ph: pancreas head, Pbt: pancreas body and tail, Mets: metastasis, L: liver, P: peritoneum, GEM: gemcitabine, GS: S-1 combined with gemcitabine, GnP: nab-paclitaxel combined with gemcitabine, PTX: paclitaxel, RT: radiation, PD: pancreaticoduodenectomy, DP: distal pancreatectomy, DP-CAR: distal pancreatectomy with en-bloc celiac axis resection, CHA: common hepatic artery, CA: celiac artery, PV: portal vein.

#### *3.4. Survival Analysis*

The MST of the entire study population was 10 months, and the one- and two-year survival rates were 39% and 12%, respectively (Figure 2). Patients who achieved PR (*n* = 99) and did not undergo CS exhibited significantly increased survival in comparison with other patients (15 vs. 7.5 months, *p* < 0.0001; Figure 2). The MST following initial treatment of patients who underwent CS (*n* = 32) was 37.2 months, and the one-, three-, and five-year survival rates were 100%, 51%, and 34%, respectively. These patients also exhibited significantly increased survival than those who achieved PR (37.2 vs. 18 months, *p* < 0.0001; Figure 2).

When long PR/SD was defined as PR/SD persisting for eight months or more, survival was significantly better among patients who underwent CS compared with those with long PR/SD who did not undergo CS (*n* = 97) (37.2 vs. 19.5 months, *p* < 0.0001).

**Figure 2.** Overall survival of all patients, patients with radiographic partial response, and patients who underwent conversion surgery. The median overall survival (OS) for the study population (solid line, *n* = 468) was 10 months. Survival was significantly better among patients with partial response (dashed line, *n* = 99) compared with other cases (*p* < 0.0001). Survival of patients who underwent conversion surgery (dotted line, *n* = 32) was significantly better than those with partial response (*p* < 0.0001). Abbreviations: CS, conversion surgery; Pts, patients; PR, partial response.

#### *3.5. Comparison between Patients with Unresectable Locally Advanced and Metastatic Disease*

Age, gender, tumor location, tumor diameter, tumor markers, pretreatment period to operation, postoperative complications, mortality, and length of hospital stay were not significantly different patients with UR-LA who underwent CS and those with UR-M who underwent CS (Table 2). Significant differences were identified in metastatic site and requirement of additional radiation therapy. There was no significant difference in survival from the time of initial treatment or from the time of CS between patients who underwent CS with UR-LA and those with UR-M (50.5 vs. 29.0 months, *p* = 0.53; 25.0 vs. 21.0 months, *p* = 0.61, respectively; Figure 3).

**Figure 3.** Overall survival of patients with unresectable locally advanced or metastatic disease who underwent conversion surgery. There was no significant difference in overall survival between patients with unresectable locally advanced (solid line, *n* = 17) and unresectable-metastatic disease (dashed line, *n* = 15) (*p* = 0.53). Abbreviations: UR-LA, unresectable locally advanced; UR-M, unresectable metastatic.

#### *3.6. Recurrence-Free Survival*

The MST from CS was 23 months, and the median recurrence-free survival time was 13 months (Figure 4).

**Figure 4.** Recurrence-free survival of patients who underwent conversion surgery. The median recurrence-free survival time of patients who underwent conversion surgery was 13 months. Abbreviations: CS, conversion surgery.

Recurrence was confirmed in 20 (62.5%) of 32 patients who underwent CS, presenting as peritoneal dissemination in seven patients, locoregional recurrence in six, liver metastasis in five, and lung metastasis in two. Recurrence within six months after CS was observed in six patients (18.8%), presenting as liver metastasis in three patients, peritoneal dissemination in two, and local recurrence in one. One of those patients received GEM, and four patients received S-1 as adjuvant chemotherapy after CS. After relapse was confirmed, two of the six patients received the same regimen as was administered for initial treatment; these patients survived 23 and 32 months after CS. Patients who suffered recurrence within six months after CS had relatively poorer prognoses than non-recurrent patients (25.5 vs. 50.5 months, *p* = 0.22). Multivariate logistic regression analyses revealed that lymph node metastasis, washing cytology positive, large tumor (>35 mm), and lack of postoperative adjuvant chemotherapy were predictive factors for early recurrence (Table 3).


**Table 3.** Predictive factor for the recurrence within six months after CS (Univariate and multivariate logistic regression analyses).

CS: conversion surgery, HR: hazard ratio, CI: confidential interval, LN: lymph node, R: residual tumor, CY: washing cytology, Tx: chemotherapy.

#### *3.7. Prognostic Factors for Overall Survival Among Patients Who Underwent Conversion Surgery*

The multivariate analysis revealed the site of pancreatic lesion and postoperative adjuvant chemotherapy to be statistically significant prognostic factors for OS among patients undergoing CS (*p* = 0.0092 and *p* < 0.0001, respectively). Other parameters, including reduction of tumor markers and Evans grading, were not significantly risk factors (Table 4).


**Table 4.** Univariate and multivariate analysis of prognostic factor of overall survival in CS group.

CS: conversion surgery, HR: hazard ratio, CI: confidential interval, LN: lymph node, R: residual tumor, CY: washing cytology, Tx: chemotherapy.

#### **4. Discussion**

Despite recent advances in diagnostic medicine, detection of pancreatic cancer while it is within the resectable stage remains a clinical challenge. According to systematic reviews, the condition is not detected until it has reached the locally advanced or metastatic stage in 30–40% and 40–50% of patients, respectively [15–17]. Thus, despite the development of chemotherapy, the prognosis of patients with UR-PDAC remains poor, with a median survival of 9.2–13.5 months and low rates of long-term survival. [5–7]

Favorable outcomes may be achieved for a certain period of time, through the use of chemo(radiation)therapy for patients with unresectable malignancies, and this treatment can be converted to surgical resection, as required. Conversion surgery represents a new therapeutic strategy which may improve short- and long-term outcomes of patients with UR-PDAC. Several articles have reported the utility of CS in such patients, as well as the positive effects on prognosis [17–26]. In the present study, the rate of CS among patients with UR-LA and UR-M was similar to that reported previously [25]. We found the long-term prognosis; one-, three-, and five-year OS rates from initial treatment; and MST were significantly better among patients with long PR/SD who did not undergo CS, although there were no significant differences in survival with relation to UR-LA or UR-M. Therefore, CS should be considered even for patients initially diagnosed with UR-M if they exhibit surgical indicators. Considering the favorable long-term survival of patients who underwent CS in the present study, our suggestion of tumor extension to the major vessels with attachment as an indication for surgery appears reasonable. However, early recurrence was observed in almost 20% of patients, in line with the findings of Wright et al., who reported that seven out of 23 patients (30.4%) with metastatic PDAC who underwent CS experienced early recurrence. Other studies have also reported early recurrence rates after conversion surgery of approximately 30% [27–29]. This would suggest that patients cannot be expected to survive longer than patients who receive non-surgical treatment, and conversion surgery may be harmful to patients because of the high risk of mortality and morbidity associated with extensive pancreatectomy. The early recurrence rate should be decreased as much as possible for patients undergoing CS [30]. Thus, although CS can prolong OS, early recurrence remains a considerable risk. Appropriate preoperative selection of patients for CS is absolutely necessary in

order to improve prognosis. The relatively strict surgical indication employed in the present study resulted in prolonged survival and a reduced incidence of early recurrence. In contrast, a review article reported that some authors recommend patients with UR-PDAC who did not experience progression after chemo(radiation) therapy should be offered surgical exploration [30]. The resectability and MST of patients in these studies who underwent CS ranged from 20% to 69% (median, 52%) and from 19.5 to 33 months (median, 21.9 months), respectively. Strict criteria may lead to lower resectability but longer OS, as a result of patient selection. Broad criteria may be associated with higher resectability but shorter OS, due to the risk of early recurrence after conversion surgery. Surgical indications for CS should be carefully decided through discussion in a multidisciplinary meeting.

To the best of our knowledge, there have been no previous studies on predictive factors of early recurrence after CS. The present study demonstrates that lymph node metastasis, positive washing cytology, large tumor size (>35 mm), and lack of postoperative adjuvant chemotherapy are significant predictive factors for early recurrence after CS. Thus, tumor size and washing cytology may be important preoperative factors which should be considered during patient selection for CS. Staging laparoscopy should be routinely performed before proceeding with CS in order to exclude patients with positive washing cytology. Metastatic site, decreased CA19-9 level, and performance status are not significant predictive factors for early recurrence. Several articles have reported that decreased CA19-9 levels after multimodal therapy represent a reliable predictive factor for resectability, OS, and DFS [21,29–35]. In most patients of the present study, CA19-9 decreased to within normal limits after multimodal treatment. Although the optimal selection criteria for surgical exploration or resection remain controversial for patients with initially UR-PDAC, it may be appropriate to base decision-making for CS on clinical response (defined by RECIST criteria) and decreased CA19-9 level after multimodal therapy [30].

Regarding pathological examination, the utility of Evans classification reflecting the extent of tumor degeneration or necrosis has been extensively studied as a prognostic factor after preoperative treatment [24,35–38]. There have been reports of the association between histopathological responses to chemo(radiation)therapy and the prognosis of patients with PDAC [24,35–38]. Chaterjee et al [36]. reported that 42 (18.8%) of 223 patients with resectable PDAC who received neoadjuvant chemotherapy were classified as Evans grade ≥III and had better survival rates than patients classified as Evans grade <III. Moreover, White et al. [37] suggested histologic response to be a useful surrogate marker for treatment efficacy, but Evans grade was not found to be a prognostic factor of CS in the present study.

The present study has some limitations which should be acknowledged. Firstly, it is a single-institute and retrospective study involving a small number of patients. All studies on this subject, to date, are retrospective studies, and so we believe that a prospective study is necessary to define the efficacy of CS. In Japan, the results of the PREP-04 trial (UMIN000017793)—a multi-institutional prospective cohort study investigating clinical outcomes of CS on patients with initially UR-PDAC— will be published in the near future. Given that only patients who responded favorably to chemo(radiation)therapy were analyzed among all patients with UR-PDAC, a selection bias exists. The development of an effective therapeutic strategy involving combined multimodal treatment with surgical resection is critical.

#### **5. Conclusions**

In conclusion, CS can provide clinical benefits, including increased survival for patients with initially UR-PDAC who have responded favorably to chemo(radiation)therapy. In addition to CS, postoperative adjuvant chemotherapy is necessary to prolong survival. It is essential that efforts are made to reduce early recurrence and to investigate surrogate markers in order to determine appropriate indications for surgery.

**Author Contributions:** Data curation, S.Y., S.H., M.K., and H.R.; methodology, M.I.; supervision, Y.M. and M.S.; writing—original draft, H.Y.; writing—review and editing, S.S., T.Y., and M.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** No funding was received for this work.

**Conflicts of Interest:** All authors declare that they have no conflict of interest. Human and animal rights: This article does not contain any studies with animals performed by any of the authors.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Systematic Review* **Systematic Review and Meta-Analysis of Surgical Treatment for Isolated Local Recurrence of Pancreatic Cancer**

**Simone Serafini <sup>1</sup> , Cosimo Sperti 1,\*, Alberto Friziero <sup>1</sup> , Alessandra Rosalba Brazzale <sup>2</sup> , Alessia Buratin <sup>3</sup> , Alberto Ponzoni <sup>4</sup> and Lucia Moletta <sup>1</sup>**


**Simple Summary:** Recurrences after primary resection of pancreatic cancer are generally treated with chemotherapy or best supportive care. Despite some reports of encouraging results after the re-resection of recurrences, the real role of surgery in this setting remains unclear. The aim of our systematic review and meta-analysis was to define the benefit of surgery in the case of isolated local recurrence. The data collected on 431 patients suggest an overall survival benefit of 29 months for patients re-operated compared to patients given medical therapies. In selected patients with recurrent pancreatic cancer, resection is safe and feasible, and may offer a survival advantage.

**Abstract:** Purpose: To perform a systematic review and meta-analysis on the outcome of surgical treatment for isolated local recurrence of pancreatic cancer. Methods: A systematic review and metaanalysis based on Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines was conducted in PubMed, Scopus, and Web of Science. Results: Six studies concerning 431 patients with recurrent pancreatic cancer met the inclusion criteria and were included in the analysis: 176 underwent redo surgery, and 255 received non-surgical treatments. Overall survival and post-recurrence survival were significantly longer in the re-resected group (ratio of means (ROM) 1.99; 95% confidence interval (CI), 1.54–2.56, *I* <sup>2</sup> = 75.89%, *p* = 0.006, and ROM = 2.05; 95% CI, 1.48–2.83, *I* <sup>2</sup> = 76.39%, *p* = 0.002, respectively) with a median overall survival benefit of 28.7 months (mean difference (MD) 28.7; 95% CI, 10.3–47.0, *I* <sup>2</sup> = 89.27%, *p* < 0.001) and median survival benefit of 15.2 months after re-resection (MD 15.2; 95% CI, 8.6–21.8, *I* <sup>2</sup> = 58.22%, *p* = 0.048). Conclusion: Resection of isolated pancreatic cancer recurrences is safe and feasible and may offer a survival benefit. Selection of patients and assessment of time and site of recurrence are mandatory.

**Keywords:** isolated local recurrence; pancreatectomy; pancreatic cancer; pancreatic remnant; recurrence; redo surgery

#### **1. Introduction**

Pancreatic ductal adenocarcinoma (PDAC) was the fourth cause of cancer-related death in the United States and in Italy in 2018, with an estimated 55.440 and 13.300 new cases diagnosed, and 44.300 and 11.463 related deaths, respectively [1,2]. Surgical resection continues to be the only chance of cure, but the 5-year survival rate remains low, ranging from 19% to 27% [3,4]. Such disappointing results are justified by the aggressive biology of pancreatic cancer and the high rate of recurrence (up to 80%) even after radical

**Citation:** Serafini, S.; Sperti, C.; Friziero, A.; Brazzale, A.R.; Buratin, A.; Ponzoni, A.; Moletta, L. Systematic Review and Meta-Analysis of Surgical Treatment for Isolated Local Recurrence of Pancreatic Cancer. *Cancers* **2021**, *13*, 1277. https://doi.org/10.3390/ cancers13061277

Academic Editor: Sohei Satoi

Received: 11 February 2021 Accepted: 10 March 2021 Published: 13 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

resections [5]. Recurrent pancreatic cancer poses a challenge for clinicians and is commonly treated with chemotherapy or best supportive care. Unlike other cancers, re-surgery for relapsing pancreatic cancer is not generally considered an option because evidence regarding its benefits is lacking, and isolated tumor recurrences amenable to resection are relatively uncommon. An isolated local recurrence (ILR) is usually defined as a tumoral recurrence localized to the posterior resection margin, the pancreatic remnant, or the locoregional lymph nodes. Some authors have recently reported encouraging results of surgical management of pancreatic cancer recurrence in selected patients [6–9].

The aim of the present study was to conduct a systematic review and meta-analysis of the outcome of redo-surgery for patients with isolated local recurrent PDAC after initial pancreatectomy.

#### **2. Materials and Methods**

#### *2.1. Study Selection*

A systematic literature search was conducted using PubMed, Scopus, and Web of Science to identify all studies published up to 30 November 2020 regarding the recurrence of pancreatic cancer after surgery. The search terms used were "pancreatic cancer/neoplasm/adenocarcinoma", "recurrence", "surgery/pancreatectomy/redo surgery/completion pancreatectomy". The articles found were used to broaden the search, and all emerging abstracts, studies, and citations were reviewed. The reference lists of all the studies considered were also screened for any other potentially relevant papers.

#### *2.2. Inclusion and Exclusion Criteria*

The following inclusion criteria were considered for the studies: (1) they reported on patients with histologically proven ILR PDAC treated surgically with curative intent, with or without (neo) adjuvant chemotherapy and/or radiotherapy; (2) they provided data on patients reoperated for recurrent PDAC after initial pancreatectomy, and their long-term outcomes; (3) they were written in English. The following exclusion criteria were considered: (1) reviews without original data or animal studies; (2) absence of individual patient data; (3) duplications; (4) lack of long-term data; and (5) in the event of successive publications by the same group, only the most detailed study was included.

Three independent reviewers (SS, ARB, AB) extracted the data using standardized data forms. All data from each eligible study were entered in a dedicated spreadsheet (Excel 2007, Microsoft Corporation®, Padua, Italy). Disagreements between the reviewers were solved by discussion and consensus. The following data were collected: title, first author, year of publication, characteristics of study population, study design, number of patients who underwent re-resection, disease-free interval (DFI), overall survival (OS), and post-recurrence survival (PRS). The articles included in this review were chosen in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [10].

#### *2.3. Terminology and Definitions*

Disease-free interval (DFI) was defined as the interval between the date of primary tumor resection and the date of recurrence. Overall survival (OS) was defined as the time between the primary tumor resection and death or latest follow-up. Post-recurrence survival (PRS) was the time interval between recurrence detection or reoperation and death or latest follow-up.

#### *2.4. Statistical Analysis*

The statistical methods of this study were reviewed by two authors (BAR and BA). Two meta-analyses were conducted in line with the Cochrane Collaboration guidelines on the Meta-analysis of Observational Studies in Epidemiology [11,12]. The first analysis focused on OS in months; the second focused on survival in months after the recurrence of PDAC. The data used for the meta-analyses are summarized in Table 1. Survival was

retrieved from the published studies as median values and ranges. Where not stated explicitly, these values were calculated from the data reported in the papers or extrapolated from the Kaplan-Meier (K-M) plots. Since all articles used in this analysis reported only the size of the study groups (without standard errors), all median survival times were converted into means and variances using a dedicated statistical algorithm [13].

The survival data were pooled for the analysis of either the mean difference (MD) or the logarithm of the ratio of means (ROM) [14]. Values of MD > 0 or ROM > 1 indicate a higher survival rate for patients who underwent re-resection. Cochran's Q statistic and the *I* 2 statistic were used to test between-study heterogeneity [15]. If the Q statistic was significant at the 0.1 level, the summary effect and corresponding 95% confidence interval (CI) were obtained with the Mantel-Haenszel random effects model [16]. For *I* <sup>2</sup> < 50%, between-study heterogeneity was judged to be low moderate; for *I* <sup>2</sup> <sup>≥</sup> 50%, it was considered substantial. The point estimate of MD and ROM was considered statistically significant when *p* was < 0.05. A cumulative meta-analysis was also run to test the stability of the pooled endpoint estimates. Publication bias was assessed visually using a funnel plot, and the number of missing studies was estimated using the trim-and-fill method [17,18]. All analyses were conducted using R version 3.5.2 [19].

**Table 1.** Studies included in the meta-analysis. OS and PFR are compared for each study (significant difference for *p* value < 0.05). Pts: number of patients; OS: overall survival; PRS: post recurrence survival; \* extrapolated from K-M plot; ˆ extrapolated from confidence interval; NR: not reported.


#### **3. Results**

The study search and selection strategy are shown in the flow chart in Figure 1. The preliminary literature search identified 1326 studies matching the initial search criteria. After screening, six studies were ultimately included in the quantitative synthesis (metaanalysis) [20–25].

It was not possible to include DFI in this analysis because this information was lacking in most studies. Likewise, due to the lack and fragmentation of precise data about the time to recurrence, it was not possible to estimate and compare conditional survival of resected and non-resected patients. Four reports [21–24] were included for OS and five for PRS [20–23,25] analysis. All six studies were retrospective and concerned a total of 431 patients with recurrent pancreatic cancer after primary pancreatic resection: 176 treated with re-resection, and 255 given non-surgical removal of recurrence. The characteristics of the selected studies are summarized in Table 1.

**Figure 1.** Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. **Figure 1.** Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.

It was not possible to include DFI in this analysis because this information was lacking in most studies. Likewise, due to the lack and fragmentation of precise data about the After surgery for recurrent PDAC, the mortality rate reported was 1.1% (2/176 patients) [20,25]; the morbidity rate ranged from 6% [25] to 33% [24].

time to recurrence, it was not possible to estimate and compare conditional survival of resected and non-resected patients. Four reports [21–24] were included for OS and five for PRS [20–23,25] analysis. All six studies were retrospective and concerned a total of 431 patients with recurrent pancreatic cancer after primary pancreatic resection: 176 treated A random-effects meta-analytical model was used for all variables. The random-effects method was always chosen because Cochran's Q statistic proved statistically significant at the *p* < 0.05 level for all meta-analyses, with a borderline situation for the analysis focusing on mean survival after recurrent PDAC, for which it was *p* = 0.048.

with re-resection, and 255 given non-surgical removal of recurrence. The characteristics of the selected studies are summarized in Table 1. After surgery for recurrent PDAC, the mortality rate reported was 1.1% (2/176 patients) [20,25]; the morbidity rate ranged from 6% [25] to 33% [24]. OS was almost twice as long in patients reoperated for ILR PDAC than in patients given non-surgical treatments (ROM 1.99; 95% CI, 1.54–2.56, *I* <sup>2</sup> = 75.89%, *p* = 0.006). The median survival benefit for patients who underwent re-resection was 28.7 months (MD 28.7; 95% CI, 10.3–47.0, *I* <sup>2</sup> = 89.27%, *p* < 0.001), as shown in Figure 2. *Cancers* **2021**, *13* 5 of 11

**Figure 2.** Forest plot for overall survival mean difference in months. RE: random-effects.

patients not re-resected (ROM = 2.05; 95% CI, 1.48–2.83, *I2* = 76.39%, *p* = 0.002): it was 15.2

**Figure 3.** Forest plot for post recurrence survival mean difference in months. RE: random-effects.

The cumulative meta-analysis demonstrated that the benefit after pancreatic re-resection settles very quickly for both OS and PRS. Funnel plots provided some evidence of

**Figure 2.** Forest plot for overall survival mean difference in months. RE: random-effects.

Median PRS was significantly longer for the group treated surgically than for the Median PRS was significantly longer for the group treated surgically than for the patients not re-resected (ROM = 2.05; 95% CI, 1.48–2.83, *I* <sup>2</sup> = 76.39%, *p* = 0.002): it was

58.22%, *p* = 0.048) (Figure 3).

15.2 months longer in the former than in the latter group (MD 15.2; 95% CI, 8.6–21.8, *I* <sup>2</sup> = 58.22%, *p* = 0.048) (Figure 3). months longer in the former than in the latter group (MD 15.2; 95% CI, 8.6–21.8, *I2* = 58.22%, *p* = 0.048) (Figure 3).

Median PRS was significantly longer for the group treated surgically than for the patients not re-resected (ROM = 2.05; 95% CI, 1.48–2.83, *I2* = 76.39%, *p* = 0.002): it was 15.2

**Figure 2.** Forest plot for overall survival mean difference in months. RE: random-effects.

*Cancers* **2021**, *13* 5 of 11

**Figure 3.** Forest plot for post recurrence survival mean difference in months. RE: random-effects. **Figure 3.** Forest plot for post recurrence survival mean difference in months. RE: random-effects.

The cumulative meta-analysis demonstrated that the benefit after pancreatic re-resection settles very quickly for both OS and PRS. Funnel plots provided some evidence of The cumulative meta-analysis demonstrated that the benefit after pancreatic reresection settles very quickly for both OS and PRS. Funnel plots provided some evidence of publication bias. One study was estimated to be missing on the left for OS, while no study seems to be missing for PRS (Figures 4 and 5). publication bias. One study was estimated to be missing on the left for OS, while no study seems to be missing for PRS (Figures 4 and 5).

**Figure 4.** Funnel plots for mean difference in overall survival. Left: original data; Right: imputa-**Figure 4.** Funnel plots for mean difference in overall survival. Left: original data; Right: imputation of missing studies.

tion of missing studies.

imputation of missing studies.

ity and mortality rates.

**4. Discussion** 

**Figure 5.** Funnel plots for mean difference in post recurrence survival. Left: original data; Right:

Pancreatic ductal adenocarcinoma is recognized as a major cause of cancer-related deaths for early metastasis, extensive invasion, and poor prognosis. At diagnosis, 50% of patients present with synchronous metastases, and further 30% present with locally advanced disease, who are not suitable for upfront surgery [1]. Moreover, despite radical resection, PDAC frequently relapses, and the clinical management of recurrences is troublesome. The well-accepted treatment for recurrent PDAC is still chemotherapy, whenever feasible. Recent studies on the surgical treatment of recurrent PDAC in selected patients reported encouraging results in terms of survival, with negligible surgical morbid-

**Figure 4.** Funnel plots for mean difference in overall survival. Left: original data; Right: imputa-

publication bias. One study was estimated to be missing on the left for OS, while no study

seems to be missing for PRS (Figures 4 and 5).

**Figure 5.** Funnel plots for mean difference in post recurrence survival. Left: original data; Right: **Figure 5.** Funnel plots for mean difference in post recurrence survival. Left: original data; Right: imputation of missing studies.

#### **4. Discussion**

imputation of missing studies.

tion of missing studies.

**4. Discussion**  Pancreatic ductal adenocarcinoma is recognized as a major cause of cancer-related deaths for early metastasis, extensive invasion, and poor prognosis. At diagnosis, 50% of patients present with synchronous metastases, and further 30% present with locally advanced disease, who are not suitable for upfront surgery [1]. Moreover, despite radical resection, PDAC frequently relapses, and the clinical management of recurrences is troublesome. The well-accepted treatment for recurrent PDAC is still chemotherapy, whenever feasible. Recent studies on the surgical treatment of recurrent PDAC in selected patients reported encouraging results in terms of survival, with negligible surgical morbid-Pancreatic ductal adenocarcinoma is recognized as a major cause of cancer-related deaths for early metastasis, extensive invasion, and poor prognosis. At diagnosis, 50% of patients present with synchronous metastases, and further 30% present with locally advanced disease, who are not suitable for upfront surgery [1]. Moreover, despite radical resection, PDAC frequently relapses, and the clinical management of recurrences is troublesome. The well-accepted treatment for recurrent PDAC is still chemotherapy, whenever feasible. Recent studies on the surgical treatment of recurrent PDAC in selected patients reported encouraging results in terms of survival, with negligible surgical morbidity and mortality rates.

ity and mortality rates. Strobel et al. [20] conducted a prospective cohort study with patients with pancreatic cancer recurrence and assessed perioperative outcome, survival, and prognostic parameters. In this series, 57 patients underwent surgery for histologically proven ILR after R0/R1 resection of PDAC. ILR was resected in 41 patients. Most resections were carried out for extrapancreatic recurrences. A pancreatic re-resection was performed in 24 patients (44%). In 19 cases, a total pancreatectomy was necessary; segmental resections were possible in only 5 patients. In 11 (20%) patients with ILR in paracaval or interaortocaval lymph nodes, a simple excision with lymphadenectomy was performed. A total of 36 (63%) patients with ILR had a recurrence in close touch with visceral arteries (SMA or the celiac trunk). In these cases resection was only attempted if the arteries were not directly involved. No data about the number of R1/R0 resection after the first operation were reported. The authors [20] also assessed the potential effect of intraoperative radiation therapy (IORT). A total of 22 patients underwent surgical resection and IORT. In 16 patients, the ILR was considered not resectable because of infiltration of the mesenteric vessels; 10 of these patients received IORT (10–15 Gray). In patients with resection of ILR, the subgroup with IORT demonstrated a shorter survival than patients without IORT (17.0 vs. 29.6 months of median survival). In contrast, patients with unresectable ILR had significantly better survival with IORT (15.1 vs. 4.3 months of median survival). They concluded that benefit of percutaneous and intraoperative radiotherapy warrants further evaluation.

Miyazaki et al. [21] reported on 284 consecutive patients with pancreatic cancer who underwent initial pancreatectomy with curative intent (R0 and R1 resection). A total of 170 patients were diagnosed with recurrent pancreatic cancer, but only 11 (16.4%) developed ILR. Two out of eleven were R1 at the time of the first operation.

Hashimoto et al. [22] retrospectively analysed the survival and pathological findings of 10 patients who developed remnant pancreatic cancer. The authors performed

a pyrosequencing assay for KRAS (codon 12) mutations and immunohistochemistry for MUC1/MUC2, and compared the histological diagnosis of the initial tumor and the remnant pancreatic cancer in the resected group. The results indicated that four cases might have developed local recurrence of the primary lesions, and the other four cases might have developed new primary lesions. It is important to point out that only one case out of ten was R1 resection at the time of the first operation.

Nakayama et al. [23] compared the survival outcomes of patients who developed isolated local vs. distant recurrence. In this subset of patients, only 3 out of 46 patients were considered R1 after the first pancreatic resection. Median survival after the recurrence was longer in the patients with ILR than in those with distant metastasis (44 vs. 13 months, *p*-value < 0.05).

Yamada et al. [24] conducted a multicenter survey of patients diagnosed with ILR in the remnant pancreas. Data from 114 patients were collected in a retrospective manner. Although their multivariate analysis could not identify any independent prognostic factors, the univariate analysis showed that excision of ILR, age (<65 years), body mass index (>20 kg/m<sup>2</sup> ), tumor dimensions (<20 mm), and distance from the pancreatic resection margin (>10 mm) were statistically significant positive prognostic factors. In this study, there was no correlation between the R1 margin at the time of first operation and distance to the pancreatic stump of ILR, which implies that the onset of a tumor in the remnant pancreas should not always be considered as a consequence of intrapancreatic colonization of the primitive cancer, as supported by Hashimoto et al. [22].

Kim et al. [25] reported on a cohort of 1610 consecutive patients with pancreatic cancer who underwent initial pancreatectomy with curative intent between January 2000 and December 2014 at Asian Medical Centre, Seoul, Korea. A total of 1346 patients were diagnosed with recurrent pancreatic cancer, but only 197 (14.6%) and 34 (2.5%) of these patients had isolated recurrence and ILR, respectively. Moreover, the authors performed a survival analysis according to the recurrence pattern. Survival after recurrence was better in patients who underwent resection of isolated recurrence in the remnant pancreas (median 28 vs. 12 months) and lung (median 36.5 vs. 9.5 months) than in those who did not undergo resection.

In a previously published systematic review by Moletta et al. [26] about the role of surgical resection for recurrent PDAC, an overall survival benefit after resection compared to non-resected patients was reported.

This raises some questions. First, is surgery for recurrent PDAC definitely worthwhile? If so, can it be proposed for all sites of recurrence? Which is the optimal treatment for recurrent PDAC: chemotherapy or surgery?

In this study, we confirmed the potential benefit of surgery, applying a quantitative analysis and statistical significance to the data previously reviewed. We analyzed six studies and compared the results of surgery vs. medical treatment (chemotherapy or best supportive care) for patients with recurrent PDAC. The results of our analysis confirmed that redo surgery for recurrent tumor offers a survival advantage for selected patients with a very low risk of perioperative mortality and an acceptable morbidity rate. The survival benefit (patients re-operated for ILR compared to patients given medical therapies) was estimated as about 29 and 15 months in terms of OS and PRS, respectively.

The resection of recurrent disease seems to be feasible and safe, and should be considered for selected patients with isolated pancreatic cancer local recurrences. Among the various sites affected, surgery for recurrences in the pancreatic remnant seems to be associated with a better outcome. Zhou et al. [8] reviewed the English literature until June 2016, collecting 19 articles on 55 patients who had complete pancreatectomy for relapsing PDAC. The 1, 3, and 5-year OS rates after second pancreatectomy were 82.2%, 49.2%, and 40.6%, respectively.

To date, only some reports and case series that investigate the benefit of surgery in isolated distant recurrence or in oligo metastatic patients have been published. Despite advances in surgical techniques, pancreatic surgery often results in a positive resection margin status (R). In particular, R0 was defined as a distance from the tumor to the closest resection margin of >1 mm, whereas R1 as a distance of ≤1 mm to the resection margin or margin involved without macroscopic involvement. R1 resection and medial/posterior margin due to perineural invasion, regional lymph node metastases, and systemic spreading at the time of surgery resulted in early recurrences. Based on today's knowledge, it is not possible to define if ILR are de novo tumors or recurrence of the primary ones. Encouraging survival rates have been reported for isolated lung metastases: the median survival time after lung metastasectomy ranged from 18.6 [9] to 47 months [27]. On the other hand, poor outcomes have been reported after the resection of liver or peritoneal metastases [28,29]. New randomized clinical trials should be conducted to better define the role of surgery in this setting.

We cannot clearly answer the question regarding the optimal treatment for recurrent PDAC because of the lack of controlled studies comparing the outcomes of patients who undergo surgery with those given chemotherapy. In the era of multi-agent systemic therapy, survival for pancreatic cancer has globally increased. However, even if many different randomized clinical trials for borderline, locally advanced, and "de novo" metastatic PDAC are published confirming the survival benefit of new regimen, in the case of recurrent PDAC, such studies are missing, and there is currently no scientific evidence supporting a specific treatment. Some authors have shown improvement in survival after treatment of recurrent PDAC with intensified regimes including FOLFIRINOX (FFN) and nab-paclitaxel plus gemcitabine (GEMNAB) compared with single agent chemotherapy. Gbolahan et al. [30] reported that administration of FFN or GEMNAB compared with single agent chemotherapy was associated with a statistically significant survival benefit, with a median OS of 14 (95% CI 9–17) vs. 8 (95% CI 6–12) months. Javed et al. [31], in a recent retrospective multi-center European study, did not report any significant differences in terms of survival between FFN/GEMNAB and any other combination of gemcitabineor 5-fluorouracil-based regimen: polichemotherapy was always superior when compared with gemcitabine monotherapy with a variable median OS of 7.9–9.9 vs. 4.9 (95% CI 4.4–5.6) months. Kawaida et al. [32] reported an objective response rate of 13.6% and a progression-free survival of 7.2 months after administration of GEMNAB with an important hematological toxicity rate of 72.7% (grade 3–4).

It is reasonable to believe that the combination of new chemotherapy regimens and surgery in fit patients could improve the outcome of recurrent PDAC. Given the heterogeneous chemotherapy regimens used, the small numbers of patients included in different studies, and the fact that most patients present with multiple sites of relapse and frequently in poor health, inevitably, the surgical option can only be offered to a few, very selected patients.

Our study has some limitations to consider. The relatively small number of studies analyzed and their heterogeneity and retrospective nature entail a significant risk of selection bias. The lack of data in some studies also prevented us from measuring disease-free interval and compare conditional survival to obtain a more accurate picture of patients' outcomes. We therefore were unable to investigate the potential role of a number of factors including patient's characteristics, baseline tumor burden and stage, neoadjuvant chemotherapy, morbidity from surgical resection and adjuvant therapy, comorbidity, and functional status on failure to receive therapy. All these aspects can lead to a selection bias that it is difficult to avoid, according to the rarity of ILR in PDAC. These important questions should be further addressed in new prospective and multicentric studies.

#### **5. Conclusions**

To our knowledge, this is the first meta-analysis comparing the outcome of patients with ILR PDAC following resection or sequential chemotherapies. In selected patients with recurrent pancreatic cancer, resection is safe and feasible, and may offer a survival advantage. Surgery should be considered as part of the multimodality management of relapsing pancreatic cancer. An accurate patient selection, considering the site and time of recurrence, and a multidisciplinary approach are essential to choose the best appropriate treatment.

**Author Contributions:** S.S. acquisition of data, analysis and interpretation of data, drafting the article, final approval; C.S. conception and design of the study, critical revision, final approval; A.F. interpretation of data, drafting the article, final approval; A.R.B. acquisition of data, analysis and interpretation of data, drafting the article, final approval; A.B. acquisition of data, analysis and interpretation of data, drafting the article, final approval; A.P., interpretation of data, drafting the article, final approval; L.M. conception and design of the study, critical revision, final approval. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Systematic Review* **Meta-Analysis of Circulating Cell-Free DNA's Role in the Prognosis of Pancreatic Cancer**

**Jelena Milin-Lazovic 1,†, Petar Madzarevic 1,†, Nina Rajovic <sup>1</sup> , Vladimir Djordjevic 2,3, Nikola Milic <sup>3</sup> , Sonja Pavlovic <sup>4</sup> , Nevena Veljkovic 5,6 , Natasa M. Milic 1,7,‡ and Dejan Radenkovic 2,3,\* ,‡**


**Simple Summary:** Pancreatic cancer is an aggressive disease with a poor prognosis. The analysis of cell-free DNA (cfDNA) for genetic abnormalities is a promising new approach for the diagnosis and prognosis of pancreatic cancer patients. In this study, we conducted a systematic review and meta-analysis of studies that reported cfDNA in pancreatic ductal adenocarcinoma (PDAC). In total, 48 studies were included in the qualitative synthesis, while 44 were assessed in the quantitative synthesis, including 3524 PDAC patients. An overall negative impact of cfDNA and *KRAS* mutations on the overall (OS) and progression free survival (PFS) (HR = 2.42, 95% CI: 1.95–2.99 and HR = 2.46, 95% CI: 2.01–3.00, respectively) were found. The performance of molecular studies to assess the presence of *KRAS* mutation by liquid biopsy may support global efforts to improve outcomes for PDAC patients.

**Abstract:** Introduction: The analysis of cell-free DNA (cfDNA) for genetic abnormalities is a promising new approach for the diagnosis and prognosis of pancreatic cancer patients. Insights into the molecular characteristics of pancreatic cancer may provide valuable information, leading to its earlier detection and the development of targeted therapies. Material and Methods: We conducted a systematic review and a meta-analysis of studies that reported cfDNA in pancreatic ductal adenocarcinoma (PDAC). The studies were considered eligible if they included patients with PDAC, if they had blood tests for cfDNA/ctDNA, and if they analyzed the prognostic value of cfDNA/ctDNA for patients' survival. The studies published before 22 October 2020 were identified through the PubMED, EM-BASE, Web of Science and Cochrane Library databases. The assessed outcomes were the overall (OS) and progression-free survival (PFS), expressed as the log hazard ratio (HR) and standard error (SE). The summary of the HR effect size was estimated by pooling the individual trial results using the Review Manager, version 5.3, Cochrane Collaboration. The heterogeneity was assessed using the Cochran Q test and I<sup>2</sup> statistic. Results: In total, 48 studies were included in the qualitative review, while 44 were assessed in the quantitative synthesis, with the total number of patients included being 3524. Overall negative impacts of cfDNA and *KRAS* mutations on OS and PFS in PDAC (HR = 2.42, 95% CI: 1.95–2.99 and HR = 2.46, 95% CI: 2.01–3.00, respectively) were found. The subgroup analysis of the locally advanced and metastatic disease presented similar results (HR = 2.51, 95% CI: 1.90–3.31). In the studies assessing the pre-treatment presence of *KRAS*, there was a moderate to high degree of

**Citation:** Milin-Lazovic, J.; Madzarevic, P.; Rajovic, N.; Djordjevic, V.; Milic, N.; Pavlovic, S.; Veljkovic, N.; Milic, N.M.; Radenkovic, D. Meta-Analysis of Circulating Cell-Free DNA's Role in the Prognosis of Pancreatic Cancer. *Cancers* **2021**, *13*, 3378. https:// doi.org/10.3390/cancers13143378

Academic Editors: Sohei Satoi and Inna Chervoneva

Received: 19 May 2021 Accepted: 23 June 2021 Published: 6 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

heterogeneity (I<sup>2</sup> = 87% and I<sup>2</sup> = 48%, for OS and PFS, respectively), which was remarkably decreased in the analysis of the studies measuring post-treatment *KRAS* (I<sup>2</sup> = 24% and I<sup>2</sup> = 0%, for OS and PFS, respectively). The patients who were *KRAS* positive before but *KRAS* negative after treatment had a better prognosis than the persistently *KRAS*-positive patients (HR = 5.30, 95% CI: 1.02–27.63). Conclusion: The assessment of *KRAS* mutation by liquid biopsy can be considered as an additional tool for the estimation of the disease course and outcome in PDAC patients.

**Keywords:** cell-free DNA; pancreatic ductal adenocarcinoma; survival; meta-analysis

#### **1. Introduction**

Pancreatic cancer is an aggressive disease with a poor prognosis. Despite the constantly evolving therapeutic and diagnostic techniques, the survival rate for pancreatic cancer still remains low compared to other malignant tumors [1]. According to the American Cancer Society (ACS) and the National Cancer Institute (NCI), the overall 5-year survival rate for pancreatic cancer is below 9% [2]. Even the small percentage of people diagnosed with the local disease (10%) experience an aberrant 5-year survival rate of 37%. The vast majority see a fate of being diagnosed at the distant stage of the disease (53%), where the survival rate is 3% [2]. Pancreatic cancer's low survival rates are attributed to late diagnosis, the lack of effective chemotherapy, and surgical limitations [3]. In 2017, there were 447,700 new cases diagnosed worldwide, and 441,082 deaths due to pancreatic cancer were recorded in the same year [4,5]. Pancreatic cancer accounts for 1.8% of all cancers, but causes 4.6% of all cancer deaths, thus resulting in it being the seventh highest cause of cancer death worldwide [2].

Cell-free DNA (cfDNA) has gained attention as a potential biomarker for a large variety of malignancies (lung, breast, liver, etc.) due to the increased levels of apoptosis, necrosis, pyroptosis, mitotic catastrophes, autophagy and phagocytosis present in cancer patients [6]. Thus, the detection of cfDNA changes in serum or plasma and the uncovering of genetic abnormalities being released from malignant tumors has been considered as promising candidate technique for cancer diagnosis through liquid biopsy [6,7]. The analysis of cell-free DNA (cfDNA) for genetic abnormalities is also a new promising approach for the diagnosis and prognosis of pancreatic cancer patients. Insights into the molecular characteristics of the pancreatic cancer may provide valuable information, leading to its earlier detection and the development of targeted therapies. The identification of a circulating biomarker for pancreatic cancer, in a non-invasive manner, is an exciting area of exploration, which may lead to personalized prognosis and therapeutic optimization from simple blood tests [8]. Previous studies have suggested that the vast majority of pancreatic ductal adenocarcinoma (PDAC) harbor mutations in the *KRAS* gene, with cfDNA mutant *KRAS* being an early marker of disease recurrence [9,10]. Tumor-derived cfDNA, known as circulating tumor DNA (ctDNA), has been the subject of extensive research. However, ctDNA's clinical usability still has not been established due to the non-standardized technique for its quantification. With the introduction of digital droplet polymerase chain reaction (ddPCR), new insights in this area have been acquired [11,12]. ddPCR's ability to aid in the determination of cfDNA and ctDNA's size and level has been shown to yield prognostic value in pancreatic cancer [13]. In this study, we performed (1) a systematic review incorporating prior studies that explored the association between cfDNA and the prognosis of patients with PDAC, and (2) a meta-analysis which quantifies the association between the presence of *KRAS* mutation and overall survival (OS) and progression-free survival (PFS) in these patients.

#### **2. Material and Methods**

A systematic review was performed in accordance with the Preferred Reporting Items for Systematic Reviews [14] and the Meta-analysis of Observational Studies in Epidemiology [15]. The standardized protocol was specifically developed for the purpose of this review, and was used by independent reviewers.

#### *2.1. Study Selection*

The publications were screened for inclusion in the systematic review in two phases, and all of the disagreements were resolved by discussion at each stage with the inclusion of a third reviewer or by consensus. Studies were included based on the following criteria: (1) studies including patients with pancreatic cancer, (2) studies with blood tests for cfDNA/ctDNA, and (3) studies analyzing the prognostic value of cfDNA/ctDNA for patients' survival results. Articles containing any of the following were excluded: (1) cfDNA/ctDNA extracted from tumor tissue; (2) studies without survival outcomes, such as OS and PFS; (3) studies lacking key data for the extraction of HR; or (4) diagnostic articles.

#### *2.2. Search Strategy*

A biostatistician with expertise in conducting systematic reviews and meta-analyses (N.M.M.) and a pancreatic cancer surgeon (D.R.) developed the search strategy. Searches of the PubMed, EMBASE, Web of Science and Cochrane Library databases until 22 October 2020 were performed for studies containing key words for cfDNA and pancreatic cancer: "cellfree DNA" or "ctDNA" or "cfDNA" or "circulating DNA" or "circulating tumor DNA" or "*KRAS*", and "pancreatic cancer" or "pancreatic carcinoma" or "pancreatic adenocarcinoma". There were no restrictions on the publication language or status. The authors of relevant studies were contacted in an attempt to obtain missing data, and to confirm the information on the study methodology and the results. The authors of relevant abstracts were contacted in order to identify eligible unpublished datasets. Reference lists of the articles that are included in the analysis were searched manually, as well as relevant reviews and editorials. Experts in the field were asked to provide information on potentially eligible studies.

#### *2.3. Article Screening and Selection*

Two reviewers (J.M.L., P.M.) independently evaluated the eligibility of all of the titles and abstracts, and performed full-text screening according to the inclusion and exclusion criteria. Disagreements were resolved by consensus (J.M.L., P.M.) or arbitration (N.M.M, D.R.).

#### *2.4. Data Abstraction and Quality Assessment*

Two reviewers (J.M.L., P.M.) independently extracted the following data: the first author's name, year of publication, country, number of patients, study design, inclusion and exclusion criteria, TNM stage, sample origin, time of the sample collected, methods of DNA detection, detection markers, and information needed to assess the articles' quality. The authors were contacted to clarify and confirm the accuracy of the abstracted data. The extraction of the survival outcome data included OS, PFS, disease-free survival (DFS), recurrence-free survival (RFS) and disease-specific survival (DSS). Hazard ratios (HR) with a corresponding 95% confidence interval (95% CI) were also obtained from the related articles.

#### *2.5. Risk of Bias*

The risk of bias in the individual studies was assessed according to the following criteria proposed by the GRADE (Grading of Recommendations, Assessment, Development and Evaluations) Working Group [16]: (1) failure to develop and apply appropriate eligibility criteria (the inclusion of a control population), (2) flawed measurements of both exposure and outcome, (3) failure to adequately control for confounding variables, and (4) incomplete follow-up. Two reviewers (J.M.L., P.M.) independently evaluated the risk of bias within and across the studies, and the overall quality of the gathered evidence. An adapted version of the Newcastle–Ottawa tool for observational studies was used [17].

#### *2.6. Statistical Analysis*

The assessed outcomes were OS and PFS expressed as the log HR and standard error (SE). For articles without explicit data for the HR and 95% CI, the logHR and SE were calculated by extracting the survival rates from Kaplan Meier curves using the WebPlotDigitizer v4.4 [18]. The HR was than estimated using a calculator formulated by Tierney et al. [19]. The number of patients at risk was extracted when available; if not, the numbers were calculated taking into account the total number of patients included in the survival analysis and selected time points accounting for the censored data [6,20–31]. In addition, if the HR data were not available, but were presented in the individual-level data, the HR with corresponding 95% CI were calculated by IBM SPSS, version 25 [32]. The summary HR effect size was estimated by pooling the individual trial results using the Review Manager, version 5.3, Cochrane Collaboration. The heterogeneity was assessed using the Cochran Q test and I<sup>2</sup> statistic. According to Higgins and Thompson [33], the heterogeneity was defined as I<sup>2</sup> > 50% or *p* value < 0.10. A random effect model was used due to presence of heterogeneity in all of the analysis [33]. The weight of each study was calculated by the inverse variance method and adjusted by effect models, which determined how much each study contributed to the pooled HR. Sensitivity analyses were performed in order to evaluate the effect of the sample origin and different survival outcome measures. A subgroup analysis was performed for locally advanced and metastatic disease. A separate forest plot was constructed for each analysis showing the HR (box), 95% CI (lines) and weight (size of box) for each trial. The diamond presented the overall effect size. The presence of publication bias was assessed by a linear regression test of the funnel plot asymmetry. *p* < 0.05 was considered to be statistically significant.

#### **3. Results**

#### *3.1. Systematic Review*

A total of 5768 potentially eligible articles were found. After duplicates were removed, 3997 titles and abstracts were screened. After reading the titles and abstracts, 3694 articles were excluded because they were not original studies, examined populations other than humans (animals, cell lines), examined other diseases, did not measure ctDNA/cfDNA, or were retracted studies, author corrections or abstracts. Of the 303 reviewed full text articles, 255 were excluded because they were not written in the English language, had no survival data, had no liquid biopsy data, were methodological studies, were ongoing clinical trials, or the full-text version of the article was not available. A total of 48 articles were selected for inclusion in the systematic review, and 44 studies were included in the meta-analysis. The flow chart presenting the steps of the study selection in detail is shown in Figure 1.

**Figure 1.** Flow chart of the study selection. **Figure 1.** Flow chart of the study selection.

The characteristics of all of the 48 publications included in the systematic review are presented in detail in Table 1. Most of the studies were conducted in China or Japan (Figure 2). The studies were published between 1996 and 2020, with a minimum sample size of 10 [20] and a maximum of 210 patients [34]. The UICC/AJCC TNM classification was given by an exact number of patients in each stage in 18 of the studies [23,25,28,30,34–48], while the UICC/AJCC TNM classification was not reported for a subgroup of patients for which the ctDNA was measured, but was instead given for a total number of patients included in the study in four studies [20,49–51]. In total, 42 studies measured the cfDNA in plasma [9,13,21,24–31,34–36,38–46,48–66], three studies measured it from serum [37,47,67], two studies measured it from blood [20,22] and one study examined the cfDNA both in serum and plasma [23]. The time of the sampling was pre-treatment in 29 studies [9,13,23,24,26,29,34,36–40,43,44,46,49,52–55,57,58,60,62–67], pre/post-treatment in 10 studies [25,27,28,30,35,41,42,45,47,59], post-treatment only in two studies [48,50], pre/post and during treatment in one study [22], and six studies did not report the time of sampling [20,21,31,51,56,61]. *KRAS* was explicitly measured in 41 studies [9,20–31,34–39,41–50,52– 55,57–64,67], the cfDNA/ctDNA total concentration was measured in four studies The characteristics of all of the 48 publications included in the systematic review are presented in detail in Table 1. Most of the studies were conducted in China or Japan (Figure 2). The studies were published between 1996 and 2020, with a minimum sample size of 10 [20] and a maximum of 210 patients [34]. The UICC/AJCC TNM classification was given by an exact number of patients in each stage in 18 of the studies [23,25,28,30,34–48], while the UICC/AJCC TNM classification was not reported for a subgroup of patients for which the ctDNA was measured, but was instead given for a total number of patients included in the study in four studies [20,49–51]. In total, 42 studies measured the cfDNA in plasma [9,13,21,24–31,34–36,38–46,48–66], three studies measured it from serum [37,47,67], two studies measured it from blood [20,22] and one study examined the cfDNA both in serum and plasma [23]. The time of the sampling was pre-treatment in 29 studies [9,13,23,24,26,29,34,36–40,43,44,46,49,52–55,57,58,60,62–67], pre/post-treatment in 10 studies [25,27,28,30,35,41,42,45,47,59], post-treatment only in two studies [48,50], pre/post and during treatment in one study [22], and six studies did not report the time of sampling [20,21,31,51,56,61]. *KRAS* was explicitly measured in 41 studies [9,20–31,34–39,41–50,52–55,57–64,67], the cfDNA/ctDNA total concentration was measured in four studies [13,52,56,65], cfDNATFx was measured in one study [66], hypermethylation was measured in one study [40], TP53 was measured in one study [62], ERBB exon 17 was measured in one study [58], and SPARC MI, UCHL1 MI, PENK M and NPTX2 MI were measured in one study [51].





DSS, disease specific survival; DFS, disease free survival; qMSP PCR, quantitative methylation specific polymerase chain reaction. \* median follow up in months.

in one study [51].

**Figure 2.** Geographical overview of the patient cases included in the meta-analyses. Data from multicenter and multicountry studies were excluded. **Figure 2.** Geographical overview of the patient cases included in the meta-analyses. Data from multicenter and multicountry studies were excluded.

A total of 44 studies used polymerase chain reaction (PCR) (ddPCR in 24 [9,23– 26,28,30,31,34,37,38,41,44–46,49,52,55,57–60,63,67]; two used restriction fragment length-PCR (RFLP-PCR) [36,54]; two used nested PCR [53,56]; two used peptide nucleic acidmediated clamping (PNA clamping) [22,47]; three used beads, emulsions, amplification and magnetics (BEAMing) [27,50,64]; one used mutant allele-specific PCR (MASA PCR) [35]; one used amplification-refractory mutation system PCR (ARMS PCR) [21]; one used quantitative methylation specific polymerase chain reaction (qMSP PCR) [51]; and an explicit PCR method was not reported in eight studies [20,29,39,40,42,46,48,66]). Next-generation sequencing (NGS) was used in three studies [61,62,65], and a bioassay was used as a primary method in one study [13]. A total of 44 studies used polymerase chain reaction (PCR) (ddPCR in 24 [9,23–26,28,30,31,34,37,38,41,44–46,49,52,55,57–60,63,67]; two used restriction fragment length-PCR (RFLP-PCR) [36,54]; two used nested PCR [53,56]; two used peptide nucleic acid-mediated clamping (PNA clamping) [22,47]; three used beads, emulsions, amplification and magnetics (BEAMing) [27,50,64]; one used mutant allele-specific PCR (MASA PCR) [35]; one used amplification-refractory mutation system PCR (ARMS PCR) [21]; oneused quantitative methylation specific polymerase chain reaction (qMSP PCR) [51]; and an explicit PCR method was not reported in eight studies [20,29,39,40,42,46,48,66]). Nextgeneration sequencing (NGS) was used in three studies [61,62,65], and a bioassay was usedas a primary method in one study [13].

[13,52,56,65], cfDNATFx was measured in one study [66], hypermethylation was measured in one study [40], TP53 was measured in one study [62], ERBB exon 17 was measured in one study [58], and SPARC MI, UCHL1 MI, PENK M and NPTX2 MI were measured

#### *3.2. Pre-Treatment KRAS Mutation, and Overall and Progression-Free Survival*

A meta-analysis was performed in order to assess the relationship between the presence of *KRAS* mutations in PDAC patients and OS before treatment. A total of 35 studies had OS as an outcome. Four studies were excluded from the overall HR effect size calculation due to measuring hypermethylation in ctDNA [40] or only post-treatment cfDNA, [48,50] or performing cfDNA TFx analysis [66]. Finally, 31 studies were included in the meta-analysis. The presence of pre-treatment *KRAS* mutations had significant prognostic value for OS in PDAC (HR = 2.42, 95% CI: 1.95–2.99) (Figure 3). There was a high degree of heterogeneity in the OS analysis (I<sup>2</sup> = 87%) and a significant presence of publication bias (*p* = 0.021) (Supplemental Figure S1). The sensitivity analysis, excluding two studies which examined ctDNA in serum, showed a similar HR (HR = 2.49, 95% CI: 2.00–3.10) (Supplemental Figure S2).


mental Figure S2).

*3.2. Pre-Treatment KRAS Mutation, and Overall and Progression-Free Survival* 

A meta-analysis was performed in order to assess the relationship between the presence of *KRAS* mutations in PDAC patients and OS before treatment. A total of 35 studies had OS as an outcome. Four studies were excluded from the overall HR effect size calculation due to measuring hypermethylation in ctDNA [40] or only post-treatment cfDNA, [48,50] or performing cfDNA TFx analysis [66]. Finally, 31 studies were included in the meta-analysis. The presence of pre-treatment *KRAS* mutations had significant prognostic value for OS in PDAC (HR = 2.42, 95% CI 1.95–2.99) (Figure 3). There was a high degree of heterogeneity in the OS analysis (I2 = 87%) and a significant presence of publication bias (*p* = 0.021) (Supplemental Figure S1). The sensitivity analysis, excluding two studies which examined ctDNA in serum, showed a similar HR (HR = 2.49, 95% CI: 2.00–3.10) (Supple-

**Figure 3.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in PDAC patients and OS. **Figure 3.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in PDAC patients and OS.

> A meta-analysis was performed in order to assess the relationship between the presence of *KRAS* mutations in PDAC patients and PFS before treatment. A total of 19 studies had PFS, DFS, RFS or DSS as an outcome. The presence of pre-treatment *KRAS* mutations demonstrated a significant prognostic value for PFS in PDAC patients (HR = 2.46, 95% CI: 2.01–3.00, *n* = 19) (Figure 4). There was a high degree of heterogeneity in the PFS analysis (I2 = 48%) %) and a significant presence of publication bias (*p* < 0.001) (Supplemental Figure S3). The sensitivity analysis including only PFS as an outcome resulted in a similar HR (HR = 2.27, 95% CI 1.83–2.82, *n* = 14) (Supplemental Figure S4). A meta-analysis was performed in order to assess the relationship between the presence of *KRAS* mutations in PDAC patients and PFS before treatment. A total of 19 studies had PFS, DFS, RFS or DSS as an outcome. The presence of pre-treatment *KRAS* mutations demonstrated a significant prognostic value for PFS in PDAC patients (HR = 2.46, 95% CI: 2.01–3.00, *n* = 19) (Figure 4). There was a high degree of heterogeneity in the PFS analysis (I<sup>2</sup> = 48%) %) and a significant presence of publication bias (*p* < 0.001) (Supplemental Figure S3). The sensitivity analysis including only PFS as an outcome resulted in a similar HR (HR = 2.27, 95% CI: 1.83–2.82, *n* = 14) (Supplemental Figure S4).


**Figure 4.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in PDAC patients and PFS. **Figure 4.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in PDAC patients and PFS. **Figure 4.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in PDAC patients and PFS.

#### *3.3. Post-Treatment KRAS Mutation and Overall and Progression-Free Survival 3.3. Post-Treatment KRAS Mutation and Overall and Progression-Free Survival 3.3. Post-Treatment KRAS Mutation and Overall and Progression-Free Survival*

A total of 10 studies examined ctDNA post-treatment; the presence of post-treatment *KRAS* mutations demonstrated significant prognostic value for OS in PDAC patients (HR = 3.53, 95% CI: 2.56–4.87, *n* = 10) (Figure 5). There was a low degree of heterogeneity in the OS analysis (I2 = 24%) and no publication bias (*p* = 0.186) (Supplemental Figure S5). Patients in nine studies underwent different regimes of chemotherapy; in six studies, surgery was performed, and combined radiotherapy was performed in one study. A total of 10 studies examined ctDNA post-treatment; the presence of post-treatment *KRAS* mutations demonstrated significant prognostic value for OS in PDAC patients (HR = 3.53, 95% CI: 2.56–4.87, *n* = 10) (Figure 5). There was a low degree of heterogeneity in the OS analysis (I<sup>2</sup> = 24%) and no publication bias (*p* = 0.186) (Supplemental Figure S5). Patients in nine studies underwent different regimes of chemotherapy; in six studies, surgery was performed, and combined radiotherapy was performed in one study. A total of 10 studies examined ctDNA post-treatment; the presence of post-treatment *KRAS* mutations demonstrated significant prognostic value for OS in PDAC patients (HR = 3.53, 95% CI: 2.56–4.87, *n* = 10) (Figure 5). There was a low degree of heterogeneity in the OS analysis (I2 = 24%) and no publication bias (*p* = 0.186) (Supplemental Figure S5). Patients in nine studies underwent different regimes of chemotherapy; in six studies, surgery was performed, and combined radiotherapy was performed in one study.


**Figure 5.** Forest plot presenting the relationship between the presence of *KRAS* mutations after treatment in PDAC patients and OS. **Figure 5.** Forest plot presenting the relationship between the presence of *KRAS* mutations after treatment in PDAC patients and OS. **Figure 5.** Forest plot presenting the relationship between the presence of *KRAS* mutations after treatment in PDAC patients and OS.

The presence of post-treatment *KRAS* mutations demonstrated significant prognostic value for PFS in PDAC patients (HR = 3.53, 95% CI: 2.49–4.99, *n* = 10) (Figure 6). There was no heterogeneity in the PFS analysis (I2 = 0%) and no publication bias (*p* = 0.247) (Supplemental Figure S6). Patients in nine studies underwent different regimes of chemotherapy, and in six studies surgery was performed. The presence of post-treatment *KRAS* mutations demonstrated significant prognostic value for PFS in PDAC patients (HR = 3.53, 95% CI: 2.49–4.99, *n* = 10) (Figure 6). There was no heterogeneity in the PFS analysis (I2 = 0%) and no publication bias (*p* = 0.247) (Supplemental Figure S6). Patients in nine studies underwent different regimes of chemotherapy, and in six studies surgery was performed. The presence of post-treatment *KRAS* mutations demonstrated significant prognostic value for PFS in PDAC patients (HR = 3.53, 95% CI: 2.49–4.99, *n* = 10) (Figure 6). There was no heterogeneity in the PFS analysis (I<sup>2</sup> = 0%) and no publication bias (*p* = 0.247) (Supplemental Figure S6). Patients in nine studies underwent different regimes of chemotherapy, and in six studies surgery was performed.


**Figure 6.** Forest plot presenting the relationship between the presence of *KRAS* mutations after treatment in PDAC patients and PFS. **Figure 6.** Forest plot presenting the relationship between the presence of *KRAS* mutations after treatment in PDAC patients and PFS. **Figure 6.** Forest plot presenting the relationship between the presence of *KRAS* mutations after treatment in PDAC patients and PFS.

> Changes in cfDNA positivity during the treatment with PFS as an outcome were examined in three studies. The responders (patients who were *KRAS* positive before treatment and *KRAS* negative after treatment) had a better prognosis than the non-responders (patients who were *KRAS* positive before treatment and remained *KRAS* positive after the treatment) (HR = 5.30, 95% CI: 1.02–27.63, *n* = 3) (Supplemental Figure S7). Changes in cfDNA positivity during the treatment with PFS as an outcome were examined in three studies. The responders (patients who were *KRAS* positive before treatment and *KRAS* negative after treatment) had a better prognosis than the non-responders (patients who were *KRAS* positive before treatment and remained *KRAS* positive after the treatment) (HR = 5.30, 95% CI: 1.02–27.63, *n* = 3) (Supplemental Figure S7). Changes in cfDNA positivity during the treatment with PFS as an outcome were examined in three studies. The responders (patients who were *KRAS* positive before treatment and *KRAS* negative after treatment) had a better prognosis than the non-responders (patients who were *KRAS* positive before treatment and remained *KRAS* positive after the treatment) (HR = 5.30, 95% CI: 1.02–27.63, *n* = 3) (Supplemental Figure S7).

#### *3.4. Analysis of the Locally-Advanced and Metastatic Disease 3.4. Analysis of the Locally-Advanced and Metastatic Disease 3.4. Analysis of the Locally-Advanced and Metastatic Disease*

A subgroup analysis of the studies examining the locally advanced and metastatic PDAC showed that *KRAS* mutations had significant prognostic value for OS (HR = 2.51, 95% CI: 1.90–3.31, *n* = 15) (Figure 7). There was a high degree of heterogeneity in the OS analysis (I2 = 79%) but no publication bias (*p* = 0.061) (Supplemental Figure S8). In the analysis examining only the metastatic disease, the effect was similar (HR = 1.90, 95% CI: 1.39–2.61, *n* = 6) (Supplemental Figure S9). A subgroup analysis of the studies examining the locally advanced and metastatic PDAC showed that *KRAS* mutations had significant prognostic value for OS (HR = 2.51, 95% CI: 1.90–3.31, *n* = 15) (Figure 7). There was a high degree of heterogeneity in the OS analysis (I<sup>2</sup> = 79%) but no publication bias (*p* = 0.061) (Supplemental Figure S8). In the analysis examining only the metastatic disease, the effect was similar (HR = 1.90, 95% CI: 1.39–2.61, *n* = 6) (Supplemental Figure S9). A subgroup analysis of the studies examining the locally advanced and metastatic PDAC showed that *KRAS* mutations had significant prognostic value for OS (HR = 2.51, 95% CI: 1.90–3.31, *n* = 15) (Figure 7). There was a high degree of heterogeneity in the OS analysis (I2 = 79%) but no publication bias (*p* = 0.061) (Supplemental Figure S8). In the analysis examining only the metastatic disease, the effect was similar (HR = 1.90, 95% CI: 1.39–2.61, *n* = 6) (Supplemental Figure S9).


**Figure 7.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in locallyadvanced and metastatic PDAC patients and OS. **Figure 7.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in locallyadvanced and metastatic PDAC patients and OS. **Figure 7.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in locallyadvanced and metastatic PDAC patients and OS.

> A subgroup analysis of the studies examining locally-advanced and metastatic PDAC showed that *KRAS* mutations demonstrated significant prognostic value for PFS (HR = 2.51, 95% CI: 1.98–3.19, *n* = 7) (Figure 8). There was not enough data to perform a separate analysis of metastatic disease with PFS as an outcome, or to test the funnel plot asymmetry (Supplemental Figure S10). A subgroup analysis of the studies examining locally-advanced and metastatic PDAC showed that *KRAS* mutations demonstrated significant prognostic value for PFS (HR = 2.51, 95% CI: 1.98–3.19, *n* = 7) (Figure 8). There was not enough data to perform a separate analysis of metastatic disease with PFS as an outcome, or to test the funnel plot asymmetry (Supplemental Figure S10). A subgroup analysis of the studies examining locally-advanced and metastatic PDAC showed that *KRAS* mutations demonstrated significant prognostic value for PFS (HR = 2.51, 95% CI: 1.98–3.19, *n* = 7) (Figure 8). There was not enough data to perform a separate analysis of metastatic disease with PFS as an outcome, or to test the funnel plot asymmetry (Supplemental Figure S10).


**Figure 8.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in locallyadvanced and metastatic PDAC patients and PFS. **Figure 8.** Forest plot presenting the relationship between the presence of *KRAS* mutations before treatment in locallyadvanced and metastatic PDAC patients and PFS.

#### **4. Discussion 4. Discussion**

In this study, we found an overall negative impact of *KRAS* mutations on OS and PFS in PDAC (HR = 2.42, 95% CI 1.95–2.99 and HR = 2. 46, 95% CI: 2.01–3.00, respectively). The subgroup analysis of locally-advanced and metastatic disease presented similar results (HR = 2.51, 95% CI: 1.90–3.31). In studies assessing the pre-treatment presence of *KRAS* mutations, there was a high degree of heterogeneity in OS (I2 = 87%) and a moderate level of heterogeneity in the PFS analysis (I2 = 48%), which was remarkably decreased in the analysis of studies measuring post-treatment *KRAS* mutations (I2 = 24% and I2 = 0%, for In this study, we found an overall negative impact of *KRAS* mutations on OS and PFS in PDAC (HR = 2.42, 95% CI: 1.95–2.99 and HR = 2. 46, 95% CI: 2.01–3.00, respectively). The subgroup analysis of locally-advanced and metastatic disease presented similar results (HR = 2.51, 95% CI: 1.90–3.31). In studies assessing the pre-treatment presence of *KRAS* mutations, there was a high degree of heterogeneity in OS (I<sup>2</sup> = 87%) and a moderate level of heterogeneity in the PFS analysis (I<sup>2</sup> = 48%), which was remarkably decreased in the analysis of studies measuring post-treatment *KRAS* mutations (I<sup>2</sup> = 24% and I<sup>2</sup> = 0%, for OS and PFS, respectively).

OS and PFS, respectively). There is a constant effort to find novel biomarkers which could improve the diagnosis, follow-up and therapeutic approaches in pancreatic cancer. The discovery that nucleic acids originating from cancer cells can be found in the peripheral circulation of cancer patients has had a major impact towards the development of non-invasive techniques, such as liquid-biopsy methodology, for the detection of tumor biomarkers. The analysis of cell-free DNA (cfDNA) for genetic abnormalities is a new promising research area for the diagnosis and prognosis of pancreatic cancer patients. CfDNA is also found in the blood of healthy individuals due to the continuous apoptosis/necrosis of hematopoietic cell line cells [6,68]. It usually consists of short fragments of less than 1000 base pairs (bp), with most being under 200bp [69]. When cell-free DNA originates from cancer cells, it is denoted as circulated tumor DNA (ctDNA). CtDNA is released into circulation primarily by the apoptosis of tumor cells and/or as a result of tumor necrosis [70,71]. Due to CtDNA's extremely low concentration (as low as 0.01% of total cfDNA) and its fragmented and short-sized nature, the detection of the mutational status of ctDNA is very There is a constant effort to find novel biomarkers which could improve the diagnosis, follow-up and therapeutic approaches in pancreatic cancer. The discovery that nucleic acids originating from cancer cells can be found in the peripheral circulation of cancer patients has had a major impact towards the development of non-invasive techniques, such as liquid-biopsy methodology, for the detection of tumor biomarkers. The analysis of cell-free DNA (cfDNA) for genetic abnormalities is a new promising research area for the diagnosis and prognosis of pancreatic cancer patients. CfDNA is also found in the blood of healthy individuals due to the continuous apoptosis/necrosis of hematopoietic cell line cells [6,68]. It usually consists of short fragments of less than 1000 base pairs (bp), with most being under 200bp [69]. When cell-free DNA originates from cancer cells, it is denoted as circulated tumor DNA (ctDNA). CtDNA is released into circulation primarily by the apoptosis of tumor cells and/or as a result of tumor necrosis [70,71]. Due to CtDNA's extremely low concentration (as low as 0.01% of total cfDNA) and its fragmented and short-sized nature, the detection of the mutational status of ctDNA is very challenging, and highly sensitive techniques have to be utilized for its detection.

challenging, and highly sensitive techniques have to be utilized for its detection. Different techniques are available for cfDNA/ctDNA detection: NGS, ddPCR, BEAMing, RFLP-PCR, and nested PCR, etc. For the mutational screening of cfDNA/ctDNA, the next-generation sequencing method (NGS) has been usually applied (both targeted and whole-genome sequencing). As the quantification of tumor-specific mutations in ctDNA has been shown to be more relevant for studying tumors, DdPCR was the most common technique used in the published studies due to its high sensitivity in the detection of rare mutations, its ability to quantify copy number variations and specific genomic loci, as well as its relatively simple workflow, in contrast to other methods [71]. Similar to the conventional PCR, this technology uses Taq-polymerase and primers/probes, but before the amplification reaction itself, the sample is divided into particles ("partitioning")—tens of thousands of droplets—and the PCR reaction takes place in each of them. Another difference from conventional or real-time (qPCR) is that it is possible to perform the direct quantification of the PCR product, without using a standard curve. The primary applications for ddPCR are rare allele detection in heterogeneous samples like liquid biopsies or FFPE samples of solid tumors, non-invasive prenatal diagnostics, viral load detection, gene ex-Different techniques are available for cfDNA/ctDNA detection: NGS, ddPCR, BEAMing, RFLP-PCR, and nested PCR, etc. For the mutational screening of cfDNA/ctDNA, the next-generation sequencing method (NGS) has been usually applied (both targeted and whole-genome sequencing). As the quantification of tumor-specific mutations in ctDNA has been shown to be more relevant for studying tumors, DdPCR was the most common technique used in the published studies due to its high sensitivity in the detection of rare mutations, its ability to quantify copy number variations and specific genomic loci, as well as its relatively simple workflow, in contrast to other methods [71]. Similar to the conventional PCR, this technology uses Taq-polymerase and primers/probes, but before the amplification reaction itself, the sample is divided into particles ("partitioning")—tens of thousands of droplets—and the PCR reaction takes place in each of them. Another difference from conventional or real-time (qPCR) is that it is possible to perform the direct quantification of the PCR product, without using a standard curve. The primary applications for ddPCR are rare allele detection in heterogeneous samples like liquid biopsies or FFPE samples of solid tumors, non-invasive prenatal diagnostics, viral load detection, gene expression and copy number variation, single cell gene expression profiling, and the validation of low-frequency mutations identified by sequencing analysis. Moreover,

epigenomic markers originating from tumor cells could be analyzed (methylation sites, circulating regulatory RNAs) [72]. A good agreement between BEAMing and ddPCR has been shown, with a kappa value of 0.91 (95% CI: 0.85–0.95) [73]. Recent advances in NGS technology have enabled similar sensitivity to the detection of ctDNA by ddPCR [74]. As each presented molecular platform has advantages and disadvantages, without evidence of a clear advantage for all of the purposes [28], the choice of platform should be determined to best meet the scientific and clinical questions being posed.

Most studies included measurements of ct/cf DNA in plasma, as plasma has been the preferred source for the extraction of circulating DNA. Even though serum contains a much higher amount (approximately a 2–24-times higher amount) of cfDNA than plasma, serum is not favored due to the possibility of contamination from white blood cells during clotting [75,76]. In this study, the sensitivity analysis excluding studies which used serum for cfDNA/ctDNA detection demonstrated similar results to those including only plasma measurements (HR = 2.49, 95% CI: 2.00–3.10).

Previous research has shown that the decrease in the levels of ctDNA during the treatment of PDAC patients may be a result of a significant reduction in the tumor burden. In contrast, the increase of the postoperative ctDNA levels may be due to a ctDNA release caused by tissue damage during surgery. Levy et al. showed that, in patients with PDAC, an endoscopic ultrasound fine-needle aspiration may be associated with increased an plasma concentration of cfDNA and the increased detection of mutant *KRAS* after the procedure [28]. Another reason for the increase in postoperative ctDNA levels may be a recurrence or tumor metastasis [47]. Lee et al. [42] suggested the importance of the post-operative analysis of ctDNA. Several of the studies included in this systematic review had pre/post treatment measurements of ctDNA, but only a few reported the survival between pre-positive/post-positive, pre-positive/post-negative, pre-negative/post-negative and pre-negative/post-positive patients. A meta-analysis of three studies that reported the survival between responders (pre-positive/post-negative) and non-responders (prepositive/post-positive) presented poorer survival for persistently positive *KRAS* patients (HR = 5.30, 95% CI: 1.02–27.63, *n* = 3). Based on the main results of this meta-analysis, in terms of their survival prognosis, PDAC patients may be grouped in two categories: those who are ctDNA positive with worse outcomes, and those who are ctDNA negative with better outcomes [42,47,59]. In cases where the ctDNA is detectable at diagnosis but becomes undetectable post-treatment, a reduction in the relapse risk is present in comparison with those in whom the ctDNA remains detectable. CtDNA can provide valuable information to determine the treatment decisions stratifying patients at low and high risk of the progression and recurrence of the disease. Prospective research should be conducted based on standardized protocols in order to evaluate further treatment strategies. It was observed previously that, in the subset of patients with resectable PDAC, ctDNA may assist the clinician in the timely detection of recurrence and the concordant introduction/addition of therapeutic measures [42,77]. It should be noted that most of the studies from this systematic review included patients with varying disease stages, thus limiting the interpretation of the prognostic role of ctDNA in resectable disease, or as a marker of disease recurrence. The data collected was utilized to determine the ways in which ctDNA's presence impacts the prognosis, rather than how specific ctDNA subtypes impact the prognosis or at what stage in the disease/treatment course these prognostic predictors are valid. Wild-type alternative and other onco-drivers present in cfDNA in specific patient cohorts (ex. *KRAS* G12C) are known to be highly actionable, allowing for precision medicine [78].

RAS genes (HRAS, *KRAS*, and NRAS) comprise the most frequently mutated oncogene family in human cancer. *KRAS* is mutated in 25% of all of the cancer cases, and is associated with poor disease prognosis [77]. Given that *KRAS* mutations are found in nearly all of the PDAC, this cancer type is arguably the most RAS-addicted cancer. Its roles in pancreatic cancer cell processes, such as increased proliferation, survival, migration and invasion, are well known [78]. An activating point mutation of the *KRAS* oncogene on codon 12 (exon 2) is the initiating event in the majority of PDAC cases (70–95%). *KRAS*G12D and

*KRAS*G12V mutations constitute about 80% of the *KRAS* mutations in PDAC [79]. For decades, *KRAS* oncoprotein was classified as undruggable cancer target [77]. According to growing evidence linking *KRAS* mutations to increased PDAC growth, the National Cancer Institute identified the targeting of *KRAS* as one of four major priorities for pancreatic cancer research. Targeted therapies and *KRAS* inhibitors appear to be very promising. A recent review investigating small-molecule *KRAS* inhibitors suggested that combining the antitumor effects from innovative new *KRAS* inhibitors like AMG510 with other agents, nanoparticles, or other auxiliary processes that can overcome the PDAC biochemical and tissue delivery issues offers hope for a new therapeutic way forward in PDAC [80].

Recently, the significance of a multigene approach based on liquid biopsy was highlighted to guide individual tailored therapy for PDAC patients. Alterations in other driver genes such as *CDKN2A*, *BRCA1/2*, *ERB2* and *NTRK*, etc. have been shown to be associated with PDAC, and they are also relevant to targeted treatments. In the recent study by Pishvaian et al. presenting 1856 patients with PDAC, 58% of the patients had molecular testing, actionable molecular alterations were identified in 26%, out of which 46 patients received a matched therapy as a second- or later-line therapy and presented a better OS [81]. In a study including 259 PDAC patients with varying disease stages, a potentially actionable mutation was detected in 29% [82], while in a study including patients with advanced PDAC, therapeutically relevant alterations were observed in 48% of the samples [83]. Given the difficulties that exist in obtaining a tumor sample in PDAC, the results of these studies highlight the importance of performing molecular profiling based on liquid biopsy, due to its simplicity and accessibility, and the importance of finding actionable early mutations in a tumor with limited therapeutic options. In addition, given that mutations may vary during the course of the disease, it is important to monitor these molecular changes [84]. With the ongoing debate regarding the use of neoadjuvant therapy in purely resectable PDAC patients, it should be also noted that ctDNA detection may play a relevant role in answering a key question: who, from these particular groups of patients, is a candidate for neoadjuvant therapy?

A strength of this study was the broad sensitive search strategy used across multiple bibliographic databases that resulted in 3997 articles screened and 48 studies included in the systematic review. The most recent meta-analysis of similar scope started with an initial set of 724 articles, with the inclusion of 18 articles due to its narrow specific search strategy [10]. The greatest number of patients (*n* = 3524) included in this analysis generated the most comprehensive meta-analysis of the assessment of the prognostic utility of cfDNA/ctDNA's in PDAC, while the meta-analysis assessing *KRAS* mutations included a total 2400 patients. In addition, in order to increase the utility of the data with were not directly shown, but were available in figures or as individual data, we used several recommended techniques to obtain HRs.

This study had several limitations, related to the clarification of liquid biopsy results in general, and those related to the proper understanding of the meta-analyses' results. The accurate interpretation of liquid biopsy results is rather challenging because of the presence of somatic mosaicism in plasma. One of the most common sources of the biological background noise of blood liquid biopsy is somatic mutation in blood cells [85]. The accumulation of somatic mutations in hematopoietic stem cells leads to their clonal expansion. This process, called clonal hematopoiesis (CH) is common in an aging healthy population [86]. Interestingly, not only mutations related to hematological malignancies, but also mutations in genes characteristic for solid tumors are detected as a result of CH. Mutations in the *KRAS* gene are also found as CH-mutations [87]. It is very important to exclude these non-tumor derived CH-mutations, in order to avoid the incorrect interpretation and inappropriate therapeutic management of solid tumors. CH mutations can be determined by performing the paired sequencing of plasma cfDNA and DNA from white blood cells. It is expected that artificial intelligence tools, such as machine learning, will enable the distinction between CH mutations and tumor-derived molecular alterations in liquid biopsy [85].

A relatively large number of studies were included, resulting in a wide range of initial tumor burdens, mixed-size patient groups and various methods of ctDNA detection, all of which contributed to increased heterogeneity. Different therapies, study designs and a range of follow up times also contributed to this high value of heterogeneity. Specific conclusions based on tumor stage, ctDNA concentration and mutations other than *KRAS* were not possible to derive. The studies included in our meta-analyses encompassed, predominantly, patients from European and Asian populations (Figure 2). Given that the misclassification of the variants coming from data that did not include dissimilar subpopulations could potentially lead to the inadequate treatments of individuals from underrepresented populations [88], the conclusions derived here should be treated cautiously. Large-scale population studies indicated that there are more significant numbers of population-specific variations than we believed previously [89,90]. Thus, the potential of ctDNA to improve the health outcomes for PDAC patients should be evaluated in the context of various populations. The results of the meta-analysis presenting the relationship between the presence of *KRAS* mutations before treatment and the survival of PDAC patients should be interpreted with caution due to the presence of significant publication bias.

#### **5. Conclusions**

The assessment of *KRAS* mutation by liquid biopsy can be considered as an additional tool for the estimation of the disease course and outcome in PDAC patients. While ddPCR was utilized in most studies to detect the *KRAS* mutations, due to greater test sensitivity, other technologies in the era of NGS may also be useful in clinical practice. The choice of the molecular platform should be determined in order to best meet the scientific and clinical questions being posed.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/cancers13143378/s1, Figure S1: Funnel plot of the meta-analysis presented in Figure 3, Figure S2: Forest plot presenting the relationship between presence of KRAS mutations before treatment in PDAC patients and OS: sensitivity analysis excluding samples from serum, Figure S3: Funnel plot of the meta-analysis presented in Figure 4, Figure S4: Forest plot presenting the relationship between presence of KRAS mutations before treatment in PDAC patients and PFS: sensitivity analysis including only PFS but not DFS and DSS, Figure S5: Funnel plot of the meta-analysis presented in Figure 5, Figure S6: Funnel plot of the meta-analysis presented in Figure 6, Figure S7: Forest plot comparing PFS between responders (patients who were KRAS positive before treatment and KRAS negative after treatment) and non-responders (patients who were KRAS positive before treatment and remained KRAS positive after the treatment), Figure S8: Funnel plot of the meta-analysis presented in Figure 7, Figure S9: Forest plot presenting the relationship between presence of KRAS mutations before treatment in metastatic PDAC patients and OS, Figure S10: Funnel plot of the meta-analysis presented in Figure 8.

**Author Contributions:** Conceptualization, J.M.-L., P.M., N.M.M., D.R.; methodology, J.M.-L., P.M., S.P., N.V., N.M.M., D.R.; software, J.M.-L., P.M., N.R., N.M.M.; validation, N.R., V.D., N.M., S.P., N.V.; formal analysis, J.M.-L., P.M., N.R., N.M., N.M.M., D.R.; investigation, J.M.-L., P.M., N.R., V.D., N.M., S.P., N.V., N.M.M., D.R.; resources, S.P., N.V., N.M.M., D.R.; data curation, J.M.-L., P.M., N.R., V.D., N.M., N.M.M.; writing—original draft preparation, J.M.-L., P.M., N.R., N.M., N.M.M., D.R.; writing—review and editing, J.M.-L., P.M., N.R., V.D., N.M., S.P., N.V., N.M.M., D.R.; visualization, J.M.-L., P.M., N.V., N.M.M.; supervision, S.P., N.V., N.M.M., D.R.; project administration, J.M.-L., P.M., N.M.M., D.R.; funding acquisition, J.M.-L., P.M., V.D., N.M.M., D.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** All of the data generated in this research are in the manuscript or its supplemental files.

**Acknowledgments:** We would like to thank our colleagues from Heliant Ltd., Belgrade, Serbia, for creating a geographical overview of patient cases included in the meta-analyses.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## **Is Laparoscopic Pancreaticoduodenectomy Feasible for Pancreatic Ductal Adenocarcinoma?**

**Chang Moo Kang 1,2,\* and Woo Jung Lee 1,2**


Received: 20 October 2020; Accepted: 14 November 2020; Published: 18 November 2020

**Simple Summary:** Pancreatic cancer is known to be one of the most lethal malignant diseases in gastrointestinal tract. Margin-negative pancreatectomy followed by postoperative adjuvant chemotherapy is essential treatment for long-term survival. Due to anatomical complexity and technical difficulty, laparoscopic pancreaticoduodenectomy is still controversial. However, with the advance of laparoscopic surgery, laparoscopic pancreatic resection of pancreatic head cancer has been carefully applied in well selected patients. The accumulating data are suggesting its technical feasibility, safety, and potential equivalent long-term oncologic outcome. In this review, the current status of laparoscopic pancreaticoduodenectomy for pancreatic head cancer is summarized. In addition, potential surgical indications and future perspectives of laparoscopic pancreaticoduodenectomy for pancreatic cancer are discussed for safe implementation in our clinical practice.

**Abstract:** Margin-negative radical pancreatectomy is the essential condition to obtain long-term survival of patients with pancreatic cancer. With the investigation for early diagnosis, introduction of potent chemotherapeutic agents, application of neoadjuvnat chemotherapy, advancement of open and laparoscopic surgical techniques, mature perioperative management, and patients' improved general conditions, survival of the resected pancreatic cancer is expected to be further improved. According to the literatures, laparoscopic pancreaticoduodenectomy (LPD) is also thought to be good alternative strategy in managing well-selected resectable pancreatic cancer. LPD with combined vascular resection is also feasible, but only expert surgeons should handle these challenging cases. LPD for pancreatic cancer should be determined based on surgeons' proficiency to fulfil the goals of the patient's safety and oncologic principles.

**Keywords:** laparoscopic; pancreaticoduodenectomy; pancreatic cancer

#### **1. Introduction**

Pancreatic cancer is one of the most lethal malignant tumors in the human gastrointestinal tract. Its overall survival is reported to be around 5%. Until now, margin-negative pancreatectomy and postoperative adjuvant chemotherapy is known as the standard treatment option for cure of the disease [1,2]. However, most pancreatic cancer patients are found in the advanced cancer stage. Only about 15% of the patients are eligible for resection and more than half of the patients usually develop local or systemic recurrence within 2 years after surgery. The 5-year survival of the patients who underwent radical pancreatectomy is known to be around 20%. Recent statistical perspectives estimated that pancreatic cancer will be one of the top 3 cancers killing humans in 2030 [3].

The pancreas is a difficult internal organ to be accessed by minimally invasive surgery. It is located in retroperitoneal space, and major vascular structures are near the pancreas. Therefore, laparoscopic exposure and dissection of the pancreas are difficult and even small breakage of tributary vessels will result in massive bleeding to obscure a clear surgical field. In addition, in the case of pancreatic head lesions, laparoscopic management of remnant pancreas and resected bile duct is still a great hurdle to overcome for safe surgical procedure [4]. However, with the advance of laparoscopic technique and experiences, laparoscopic distal pancreatectomy (LDP) is regarded as a safe and standard approach in well-selected left-sided pancreatic tumor [5], and even laparoscopic pancreaticoduodenectomy (LPD) is carefully thought to be an appropriate surgical option to treat periampullary lesions [6].

In the past, the Yonsei criteria was suggested as appropriate tumor conditions for oncologically safe laparoscopic radical distal pancreatectomy (LDP) [7]. According to our experience, the Yonsei criteria was found to be not only selection criteria for LDP, but also a clinically detectable parameter to predict long-term survival of left-sided pancreatic cancer [8]. Carefully expanding indications for LDP is quite acceptable as long as patients' safety and oncologic principles are maintained [9]. In the absence of randomized trials, uncertainty regarding the oncologic efficacy of LDP still exists. However, accumulating experience shows that LDP is associated with comparable survival, R0 resection, and use of adjuvant chemotherapy when comparing to open distal pancreatectomy (ODP) [10,11].

On the other hand, the current status of LPD has a long way to go in terms of technical and oncological safety. Unlike LDP, many surgeons are still within their learning curve period. Nickel et al. [12] performed a meta-analysis of randomized controlled trials comparing LPD and open pancreaticoduodenectomy (OPD) [13–15]. They concluded that at the current level of evidence, LPD shows no advantage over OPD except lower estimated intraoperative blood loss. Even though three currently available randomized control trials (RCT) were included for meta-analysis, lack of blinding of the patients and personnel assessing the main outcomes, and the learning curve issue should be considered when interpreting the results. Especially, pancreatic head cancer requires not only for skillful laparoscopic techniques but also wisdom to select appropriate patients for safe and effective margin-negative LPD, because curative resection is known to be basis for long-term survival of pancreatic cancer. Due to aggressive tumor biology and anatomical intimacy between the pancreas and major vascular structures, LPD for pancreatic cancer should be performed by expert surgeons who have already overcome of their learning-curve period [16–18]. It should be the last stage of LPD application in periampullary cancers after acquisition of full technical maturation.

It might be too early to generalize the potential oncological role of LPD in managing pancreatic head cancer, however several emerging articles from expert surgeons are providing future insight that LPD can be safe and effective in well-selected pancreatic cancer patients [19–21]. In fact, recent NCCN guideline version 1. 2020 pancreatic adenocarcinoma recommended surgical treatment by laparotomy or minimally invasive surgery as treatment for resectable pancreatic cancer [22]. LPD is no longer just a debatable issue in treating pancreatic cancer, but already regarded as one of the recommendable options in clinical oncology of pancreatic cancer. However, the articles presented as the basis for such a guideline were thought to be limited; they were published a long time ago, and were not focused on pancreatic cancer [23,24]. More detailed concerns about indications and understanding the current status of LPD in treating pancreatic cancer will be a strong background to increase the justification of clinical practice of LPD for pancreatic cancer.

#### **2. Rationale of Minimally Invasive Pancreaticoduodenectomy for Pancreatic Cancer**

#### *2.1. Technical Feasibility*

#### 2.1.1. Surgical Extent

Lymph node (LN) involvement is analyzed to be a very important prognostic factor in pancreatic cancer. Indeed, more than half of the resected pancreatic cancer showed LN metastasis and poor survival outcomes [25]. Then, for potentially clearing metastatic LNs, can extended pancreaticoduodenectomy (PD), clearing LNs around celiac, superior mesenteric artery (SMA), paraaortic, and hepatoduodenal ligament, as well as nerve plexus dissection around celiac, hepatic artery, and SMA, improve oncologic

outcome in patients with resectable pancreatic cancer (Table 1)? Due to frequent recurrence and poor long-term oncologic outcomes following curative resection of pancreatic cancer, extended dissection was once advocated in surgical management of pancreatic cancer [26–28]. However, these are all retrospective observational studies, and selection bias should be considered when interpretating surgical outcomes. To optimize the surgical extent of PD in treating pancreatic cancer, researchers performed several important prospective randomized control studies to determine the optimal extent of surgical resection in treating resectable pancreatic cancer (Table 2).

**Table 1.** Topographic differences in extent of lymph node dissection between standard pancreaticoduodenectomy (PD) and extended PD.


\* Mentioned in more than 3 RCTs [29]; HA, hepatic artery; SMA, superior mesenteric artery.

**Table 2.** Literature review of randomized control trials (RCTs) comparing standard PD and extended PD in treating pancreatic cancer.


STD, standard dissection; EXT, extended dissection; EBL, estimated blood loss; LN, lymph node; LOH, length of hospital stay; POPF, postoperative pancreatic fistula; DGE, delayed gastric emptying; Mx, mortality; 5YOS, 5-year overall survival; \*, *p* < 0.05; √ , median survival time; H , incidence of transfusion 5% vs. 23%, *p* = 0.01.

When looking at the number of retrieved LNs, extended PD shows a higher number of retrieved LNs than standard PD (Table 2). However, a recent meta-analysis [29] to review oncologic outcomes of five randomized controlled trials comparing extended and standard lymphadenectomy in patients with PD for pancreatic cancer demonstrated that PD with standard dissection was safe (shorter operation time, less transfusion, less overall postoperative complications, and similar R0) and showed similar long-term survival outcomes to that of extended dissection (HR = 1.01 [95% CI: 0.77–1.34], *p* = 0.923). It is known that intraoperative transfusion [35] and postoperative complications [36,37] have an adverse impact on survival of resected pancreatic cancer, suggesting potential benefit of standard dissection in treating pancreatic cancer. Although the number of retrieved lymph nodes is reported to be one of the important prognostic factors in treating pancreatic cancer [38,39], recently lymph node ratio, and number of positive lymph nodes, "NOT total nodes examined", were associated with overall survival in resected pancreatic cancer [40].

In summary, standard PD is not inferior and has comparable oncologic outcomes in treating resectable pancreatic cancer. Therefore, PD with routine extended lymph node dissection is not recommended due to higher morbidity and comparable long-term survival. According to the current technical availability of LPD, at least the extent of standard dissection is thought to be well achieved by current laparoscopic surgical technique (Figure 1). *Cancers* **2020**, *12*, x 4 of 20

**Figure 1.** Laparoscopic view after resection of pancreatic head cancer*.* BD; bile duct, PV; portal vein, CHA; common hepatic artery. P; pancreas, SV; splenic vein, SMV; superior mesenteric vein, SMA; superior mesenteric artery, LRV; left renal vein, IVC; inferior vena cava, gastroduodenal artery stump (white arrows), pancreatic duct (thick white arrow). **Figure 1.** Laparoscopic view after resection of pancreatic head cancer. BD; bile duct, PV; portal vein, CHA; common hepatic artery. P; pancreas, SV; splenic vein, SMV; superior mesenteric vein, SMA; superior mesenteric artery, LRV; left renal vein, IVC; inferior vena cava, gastroduodenal artery stump (white arrows), pancreatic duct (thick white arrow).

#### 2.1.2. Retroperitoneal Margin 2.1.2. Retroperitoneal Margin

Cancer cell involvement in the resection margin is known to be associated with early tumor recurrence and poor long-term oncologic outcomes [41]. The pancreatic neck and head wrap the superior mesenteric vein (SMV)/portal vein (PV), and the uncinate process of the pancreas is elongated from the pancreatic head and extends behind the SMV/PV to contact the right lateral aspect of the SMA. There are abundant lymphatics and neural tissues around this area. Surgeons need to consider not only the pancreatic neck margin, but also the retroperitoneal margin, the so called SMA lateral margin, when performing radical PD for pancreatic cancer, because pancreatic cancer cells can invade and infiltrate along the nerve tissue in the pancreas toward major arterial systems (the SMA, common hepatic artery, and celiac axis) [42–44]. Therefore, even in case of pancreatic cancer that is very separated from the SMA, pancreatic cancer cells can invade SMA through this nerve tissue. The oncologic significance of retroperitoneal margin clearance has been reported [45,46]. Butler et al. [47] recently performed systemic review on the oncologic role of periadventitial dissection of SMA in affecting margin status after PD for pancreatic cancer. According to this review, it was suggested that positive margin was associated with decreased survival and SMA margin involvement was the most often, which ranged 15–45% of resected pancreatic cancer. Surgical dissection of nerve tissue (retroperitoneal margin) around SMA is not that simple. Not Cancer cell involvement in the resection margin is known to be associated with early tumor recurrence and poor long-term oncologic outcomes [41]. The pancreatic neck and head wrap the superior mesenteric vein (SMV)/portal vein (PV), and the uncinate process of the pancreas is elongated from the pancreatic head and extends behind the SMV/PV to contact the right lateral aspect of the SMA. There are abundant lymphatics and neural tissues around this area. Surgeons need to consider not only the pancreatic neck margin, but also the retroperitoneal margin, the so called SMA lateral margin, when performing radical PD for pancreatic cancer, because pancreatic cancer cells can invade and infiltrate along the nerve tissue in the pancreas toward major arterial systems (the SMA, common hepatic artery, and celiac axis) [42–44]. Therefore, even in case of pancreatic cancer that is very separated from the SMA, pancreatic cancer cells can invade SMA through this nerve tissue. The oncologic significance of retroperitoneal margin clearance has been reported [45,46]. Butler et al. [47] recently performed systemic review on the oncologic role of periadventitial dissection of SMA in affecting margin status after PD for pancreatic cancer. According to this review, it was suggested that positive margin was associated with decreased survival and SMA margin involvement was the most often, which ranged 15–45% of resected pancreatic cancer.

uncommonly, it is usually difficult to exactly differentiate from pancreatic uncinated process and SMA due to abundant lymphatic tissue, inflammatory changes associated with pancreatitis, and neural tissue around SMA. These tissues are all intermingled altogether. In addition, it is difficult to expose the right lateral aspect of SMA because this area is behind the SMV-SV confluence. Lastly, there are abundant lymphovascular structures and even small breakage of tributary vessels around SMA and SMV usually give rise massive bleeding to prevent from further safe dissection for marginnegative resection. Therefore, considering oncologic significance of retroperitoneal margin in treating pancreatic Surgical dissection of nerve tissue (retroperitoneal margin) around SMA is not that simple. Not uncommonly, it is usually difficult to exactly differentiate from pancreatic uncinated process and SMA due to abundant lymphatic tissue, inflammatory changes associated with pancreatitis, and neural tissue around SMA. These tissues are all intermingled altogether. In addition, it is difficult to expose the right lateral aspect of SMA because this area is behind the SMV-SV confluence. Lastly, there are abundant lymphovascular structures and even small breakage of tributary vessels around SMA and SMV usually give rise massive bleeding to prevent from further safe dissection for margin-negative resection.

cancer, surgical design and plan to obtain cancer-free retroperitoneal margin is very important. There

Therefore, considering oncologic significance of retroperitoneal margin in treating pancreatic cancer, surgical design and plan to obtain cancer-free retroperitoneal margin is very important. There are several surgical techniques to secure retroperitoneal margin during LPD (Table 3). Rho et al. [48] introduced the potential application of indocyanine green (ICG) to facilitate the securement of the SMA lateral margin in laparoscopic PD, which is based on the idea that the pancreas is a well perfused organ and near infra-red light can detect ICG location through fluorescent illumination. When ICG is accumulated in the pancreatic head, near-infrared light can show visual differentiation between the uncinated process and surrounding soft tissues along the SMA (Figure 2), helping surgeons to obtain a negative retroperitoneal margin. The impact of long-term oncologic outcome of ICG-based SMA lateral border dissection remains to be investigated further, however it is thought to be a useful technique when performing LPD in resectable pancreatic cancer. Nagakawa et al. [49] introduced the technique to expose the inferior pancreaticoduodenal artery and SMA lateral border by proximal-dorsal jejunal vein (PDPV) pre-isolation, which is thought to be one of the approaches for standard PD. Moreover, Morales et al. [50] and Zimmitti et al. [51] demonstrate the surgical method to obtain the clear SMA margin by applying the SMA artery first approach. Despite requiring advanced laparoscopic skills, these methods are thought to be useful for margin-negative resection in selected cases of relatively advanced pancreatic cancer. Kuroki et al. [52] also introduced the concept of the pancreas-hanging maneuver by Penrose drain in managing SMA margin during LDP, however, they did not describe the R0 resection rate among the patients with this technique, and no patients with pancreatic head cancer were involved. *Cancers* **2020**, *12*, x 5 of 20 organ and near infra-red light can detect ICG location through fluorescent illumination. When ICG


**Table 3.** Several approaches to obtain clear SMA lateral margin. is accumulated in the pancreatic head, near-infrared light can show visual differentiation between the uncinated process and surrounding soft tissues along the SMA (Figure 2), helping surgeons to
